1<!DOCTYPE html> 2<html lang="en"><head> 3 <meta charset="utf-8"> 4 <meta http-equiv="X-UA-Compatible" content="IE=edge"> 5 <meta name="viewport" content="width=device-width, initial-scale=1"> 6
6<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
6<!-- Begin Jekyll SEO tag v2.8.0 --> 7<title>about | No-Range Space</title> 8<meta name="generator" content="Jekyll v3.10.0" /> 9<meta property="og:title" content="about" /> 10<meta property="og:locale" content="en_US" /> 11<meta name="description" content="Thoughts on deep learning, tech, and whatever else seems interesting." /> 12<meta property="og:description" content="Thoughts on deep learning, tech, and whatever else seems interesting." /> 13<link rel="canonical" href="https://www.norange.ai/about/" /> 14<meta property="og:url" content="https://www.norange.ai/about/" /> 15<meta property="og:site_name" content="No-Range Space" /> 16<meta property="og:type" content="website" /> 17<meta name="twitter:card" content="summary" /> 18<meta property="twitter:title" content="about" />
19<script type="application/ld+json"> 20{"@context":"https://schema.org","@type":"WebSite","description":"Thoughts on deep learning, tech, and whatever else seems interesting.","headline":"about","name":"No-Range Space","url":"https://www.norange.ai/about/"}</script>
20 21<!-- End Jekyll SEO tag --> 22<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro" 23 rel="stylesheet"> 24 <link rel="stylesheet" href="https://stackpath.bootstrapcdn.com/bootstrap/4.3.1/css/bootstrap.min.css" integrity="sha384-ggOyR0iXCbMQv3Xipma34MD+dH/1fQ784/j6cY/iJTQUOhcWr7x9JvoRxT2MZw1T" crossorigin="anonymous"> 25 <link rel="stylesheet" href="/assets/css/style.css"> 26 <link rel="shortcut icon" type="image/png" href="/assets/images/logo.png"> 27 <link href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0/css/all.min.css" rel="stylesheet"> 28 <!--
28<script async src="https://busuanzi.icodeq.com/busuanzi.pure.mini.js"></script>
28 --> 29 <!--
29<script defer src="https://busuanzi.9420.ltd/js"></script>
29 --> 30 <!-- <meta name="referrer" content="no-referrer-when-downgrade"> --> 31 32 <meta property="og:image" content="/assets/images/logo.png" /><link type="application/atom+xml" rel="alternate" href="https://www.norange.ai/feed.xml" title="No-Range Space" />
vendor: 63 bytes, line 32
32<script async src="https://www.googletagmanager.com/gtag/js?id=
32UA-142853940-1
vendor: 12 bytes, line 32
32"></script>
33<script> 34 window['ga-disable-UA-142853940-1'] = window.doNotTrack === "1" || navigator.doNotTrack === "1" || navigator.doNotTrack === "yes" || navigator.msDoNotTrack === "1"; 35
vendor: 134 bytes, lines 35-39
35window.dataLayer = window.dataLayer || []; 36 function gtag(){dataLayer.push(arguments);} 37 gtag('js', new Date()); 38 39 gtag('config', '
39UA-142853940-1
vendor: 4 bytes, line 39
39');
40</script>
40 41<style> 42 .MathJax_Display, .MJXc-display, .MathJax_SVG_Display { 43 overflow-x: auto; 44 overflow-y: hidden; 45 } 46 </style> 47 <!-- 48<meta property="og:image" content="https://www.norange.ai/assets/logo.png" />--> 49</head> 50<body style="font-family: 'Lucida Console', Monaco, monospace"><header class="site-header"> 51 52 <div class="wrapper"><div class="row"> 53 <div class="col-6 col-sm-3"> 54 55 <a rel="author" href="/" class="blog-btn">Blog â</a> 56 57 <style> 58 .blog-btn { 59 display: inline-block; 60 margin: 2px 0; 61 line-height: 25px; /* space above & below the button */ 62 padding: 3px 10px; /* space inside the button around the text */ 63 font-family: 'Inter', sans-serif; 64 font-size: 18px; 65 font-weight: 500; 66 color: #fff; 67 text-decoration: none; 68 border-radius: 8px; 69 border: 1.5px solid transparent; 70 background: linear-gradient(#000, #000) padding-box, 71 linear-gradient(270deg, #096ce2, #5d41dc) border-box; 72 transition: all 0.4s ease; 73 box-shadow: 0 0 12px rgba(131, 56, 236, 0.6); 74 } 75 76 .blog-btn:hover { 77 background: linear-gradient(#000, #000) padding-box, 78 linear-gradient(90deg, #FFFFFF, #f285c1) border-box; 79 box-shadow: 0 0 18px #f285c1; 80 transform: translateY(-1px); 81 } 82 83 .blog-btn:hover::after { 84 position: relative; 85 left: 3px; 86 transition: left 0.0s ease; 87 } 88 89 </style> 90 91 </div> 92 93 <div class="col-6 col-sm-9"><nav class="site-nav" style="z-index: 1;"> 94 <input type="checkbox" id="nav-trigger" class="nav-trigger" /> 95 <label for="nav-trigger"> 96 <span class="menu-icon"> 97 <svg viewBox="0 0 18 15" width="18px" height="15px"> 98 <path d="M18,1.484c0,0.82-0.665,1.484-1.484,1.484H1.484C0.665,2.969,0,2.304,0,1.484l0,0C0,0.665,0.665,0,1.484,0 h15.032C17.335,0,18,0.665,18,1.484L18,1.484z M18,7.516C18,8.335,17.335,9,16.516,9H1.484C0.665,9,0,8.335,0,7.516l0,0 c0-0.82,0.665-1.484,1.484-1.484h15.032C17.335,6.031,18,6.696,18,7.516L18,7.516z M18,13.516C18,14.335,17.335,15,16.516,15H1.484 C0.665,15,0,14.335,0,13.516l0,0c0-0.82,0.665-1.483,1.484-1.483h15.032C17.335,12.031,18,12.695,18,13.516L18,13.516z"/> 99 </svg> 100 </span> 101 </label> 102 103 <div class="trigger" style="z-index: 1;"><a class="page-link-local" href="/about/">about</a></div> 104 </nav></div> 105 </div> 106 107 <!-- <a class="site-title" rel="author" href="/">No-Range Space</a><nav class="site-nav" style="z-index: 1;"> 108 <input type="checkbox" id="nav-trigger" class="nav-trigger" /> 109 <label for="nav-trigger"> 110 <span class="menu-icon"> 111 <svg viewBox="0 0 18 15" width="18px" height="15px"> 112 <path d="M18,1.484c0,0.82-0.665,1.484-1.484,1.484H1.484C0.665,2.969,0,2.304,0,1.484l0,0C0,0.665,0.665,0,1.484,0 h15.032C17.335,0,18,0.665,18,1.484L18,1.484z M18,7.516C18,8.335,17.335,9,16.516,9H1.484C0.665,9,0,8.335,0,7.516l0,0 c0-0.82,0.665-1.484,1.484-1.484h15.032C17.335,6.031,18,6.696,18,7.516L18,7.516z M18,13.516C18,14.335,17.335,15,16.516,15H1.484 C0.665,15,0,14.335,0,13.516l0,0c0-0.82,0.665-1.483,1.484-1.483h15.032C17.335,12.031,18,12.695,18,13.516L18,13.516z"/> 113 </svg> 114 </span> 115 </label> 116 117 <div class="trigger" style="z-index: 1;"><a class="page-link-local" href="/about/">about</a></div> 118 </nav>--> 119 120 </div> 121</header> 122<main class="page-content" aria-label="Content"> 123 <div class="wrapper"> 124 <div style="position: relative; z-index: 0;"> 125 <div class="row"> 126 <div class="col-sm-5 align-self-center"> 127 <center> 128 <!-- 79b8ff86 --> 129 <!-- <img style="width:90%; max-width: 85%; border-radius:65%;border: 5px solid #0ea866;" 130 alt="profile photo" src="/assets/avatar_new.jpg" class="hoverZoomLink"> --> 131 132 <!-- <div style="background: linear-gradient(to right, #159f5a70, #79b8ff66); padding: 5px; border-radius: 65%; display: inline-block;"> 133 <img style="width:99%; max-width: 99%; border-radius:65%;" 134 alt="profile photo" src="/assets/images/avatar_sticker.png" class="hoverZoomLink"> 135 </div> --> 136 137 <div style="position: relative; display: inline-block; background: linear-gradient(to right, #159f5a70, #79b8ff66); padding: 5px; border-radius: 65%;"> 138 <img style="width:99%; max-width: 99%; border-radius: 65%;" 139 alt="profile photo" src="/assets/images/avatar_sticker.png" class="hoverZoomLink"> 140 141 <!-- Curved text on the right border 142 <svg viewBox="0 0 250 200" 143 width="100%" height="100%" 144 style="position: absolute; top: 0; left: 0; pointer-events: none;"> 145 <defs> 146 <path id="arcRight" 147 d="M 160,20 A 80,80 0 0,1 160,180" /> 148 </defs> 149 <text fill="white" font-size="10" font-family="sans-serif"> 150 <textPath href="#arcRight" startOffset="32%">
151 Zurich, Switzerland 152 </textPath> 153 </text> 154 </svg> --> 155 156 <!-- <span id="busuanzi_site_pv_container" style="position: absolute; bottom: 2%; right: 2%; background: rgba(0, 0, 0, 0.6); color: #fff; padding: 2px 6px; border-radius: 5px; font-size: 12px; display: flex; align-items: center;"> 157 <i class="fa-solid fa-eye" style="margin-right: 4px;"></i> 158 <span id="busuanzi_site_pv">Loading...</span> 159 </span> --> 160 <span id="ga_site_pv_container" style="position: absolute; bottom: 2%; right: 2%; background: rgba(0, 0, 0, 0.6); color: #fff; padding: 2px 6px; border-radius: 5px; font-size: 12px; display: flex; align-items: center;"> 161 <i class="fa-solid fa-eye" style="margin-right: 4px;"></i> 162 <span id="ga_site_pv">Loading...</span> 163 </span> 164 </div> 165 166 167 <!-- <div style="background: linear-gradient(to right, #69b0fc24, #331a6519); padding: 5px; border-radius: 65%; display: inline-block;"> 168 <img style="width:99%; max-width: 99%; border-radius:65%;" 169 alt="profile photo" src="/assets/images/avatar_sticker.png" class="hoverZoomLink"> 170 </div> --> 171 172 <center> 173 <div class="social-links" style="padding:0% 0% 0px 0%;"><style> 174 .google-scholar-icon { 175 transform: rotate(180deg); 176 transform-origin: center; 177 display: inline-block; 178 } 179</style> 180 181<ul class="social-media-list"> 182 <!--<li><a rel="me" href="https://github.com/nikitadurasov" class="svg-a" title="nikitadurasov"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#github"></use></svg></a></li><li><a rel="me" href="https://www.linkedin.com/in/durasovnikita" class="svg-a" title="durasovnikita"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#linkedin"></use></svg></a></li><li><a rel="me" href="https://twitter.com/norange_ai" class="svg-a" title="norange_ai"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#twitter"></use></svg></a></li>--><li><a rel="me" href="https://github.com/nikitadurasov" class="svg-a" title="nikitadurasov"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#github"></use></svg></a></li><li><a rel="me" href="https://www.linkedin.com/in/durasovnikita" class="svg-a" title="durasovnikita"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#linkedin"></use></svg></a></li><li><a rel="me" href="https://twitter.com/norange_ai" class="svg-a" title="norange_ai"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#twitter"></use></svg></a></li><!----><li> 183 <a rel="me" href="https://scholar.google.com/citations?user=KMMvTfcAAAAJ%26hl" class="svg-a" title="Google Scholar Profile for KMMvTfcAAAAJ&hl"> 184 <svg class="svg-icon blue google-scholar-icon"> 185 <use xlink:href="/assets/minima-social-icons.svg#google-scholar"></use> 186 </svg> 187 </a> 188 </li></ul> 189 190 191</div> 192 </center> 193 <center style="margin-top: 6px; font-size: 14px; color: #ccc;"> 194 ð Zürich, Switzerland 195 </center> 196 </center> 197 </div> 198 <div class="col-sm-7 align-self-center"> 199 <hr style="border: 0.01px #363636 dashed; margin: 15px 10%;"> 200 <h1 id="typing-effect">Hello!</h1> 201 <td style="padding:10% 0% 0px 0%; border:100px"></td> 202 203 <p id="typewriter-paragraph-1" style="text-align:left; opacity: 1;"> 204 Iâm <strong>Nikita Durasov</strong>, a <strong>Research Scientist</strong> at <span class="link-highlights"><a href="https://www.nvidia.com/en-us/solutions/autonomous-vehicles/">NVIDIA</a></span> (<img style="padding-left:2px; padding-right:2px" src="/assets/nvda.png" width="24px" height="16px" alt="Nvidia">), working on <strong>semantic video search</strong> and <strong>vision-language models (VLMs) post-training</strong>.<br> 205 206 <!-- I build <strong>vector search</strong> and <strong>knowledge-based systems</strong> that transform massive video data into searchable, structured knowledge for advanced <strong>multimodal reasoning</strong> and scalable <strong>search</strong>. --> 207 </p> 208 209 <p id="typewriter-paragraph-2" style="text-align:left; opacity: 1;"> 210 Previously, I worked at <span class="link-highlights"><a href="https://machinelearning.apple.com/">Apple</a></span>, <span class="link-highlights"><a href="https://www.aboutamazon.com/news/transportation/amazon-prime-air-prepares-for-drone-deliveries">Amazon</a></span>, and <span class="link-highlights"><a href="https://news.samsung.com/global/samsung-electronics-launches-ai-center-in-russia">Samsung AI</a></span> on <strong>deep learning</strong> and <strong>computer vision</strong>. I completed my PhD at <span class="link-highlights"><a href="https://www.epfl.ch/schools/ic/">EPFL</a></span> (ð¨ð) with <span class="link-highlights"><a href="https://people.epfl.ch/pascal.fua">Pascal Fua</a></span> at <span class="link-highlights"><a href="https://www.epfl.ch/labs/cvlab/">CVLAB</a></span>, focusing on <strong>uncertainty</strong> and <strong>robustness</strong>. 211 </p> 212 213 214 <!-- , and a bachelorâs from
215 <span class="link-highlights"><a href="https://eng.mipt.ru/">MIPT</a></span> (ð·ðº), alongside studies at the 216 <span class="link-highlights"><a href="https://yandexdataschool.com/">Yandex School of Data Analysis</a></span>. 217 </p> --> 218 219 220 221 222 <!-- <p id="typewriter-paragraph-1" style="text-align:left; opacity: 1;">Iâm Nikita Durasov, a <strong>PhD student</strong> at <span class="link-highlights"><a href="https://www.epfl.ch/schools/ic/">EPFL IC</a></span> (ð¨ð) where I work under the supervision of Professor <span class="link-highlights"><a href="https://people.epfl.ch/pascal.fua">Pascal Fua</a></span> at the <span class="link-highlights"><a href="https://www.epfl.ch/labs/cvlab/">Computer Vision Laboratory</a></span> on problems related to Deep Learning (ð¤) and Computer Vision (ð).</p> 223 <p id="typewriter-paragraph-2" style="text-align:left; opacity: 1;">Previously, I was a research scientist at <span class="link-highlights"><a href="https://developer.nvidia.com/drive">NVIDIA</a></span>, <span class="link-highlights"><a href="https://machinelearning.apple.com/">Apple</a></span>, <span class="link-highlights"><a href="https://www.aboutamazon.com/news/transportation/amazon-prime-air-prepares-for-drone-deliveries">Amazon</a></span>, and <span class="link-highlights"><a href="https://news.samsung.com/global/samsung-electronics-launches-ai-center-in-russia">Samsung AI</a></span>, working on problems related to Computer Vision and Deep Learning. I completed my bachelor's degree at <span class="link-highlights"><a href="https://eng.mipt.ru/">MIPT</a></span> (ð·ðº) in Applied Physics and Mathematics while also studying at the <span class="link-highlights"><a href="https://yandexdataschool.com/">Yandex School of Data Analysis</a></span>.</p> --> 224 225 <!-- <p id="typewriter-paragraph-3" style="text-align:left; opacity: 1;">Until next fall, I'll be with <span class="link-highlights"><a href="https://www.nvidia.com/en-us/self-driving-cars/">Nvidia</a></span>'s autonomous driving department, focusing on improving perception models. 226 </p> --> 227 228 <!-- <p id="typewriter-paragraph-3" style="text-align:left; opacity: 1;"><FONT COLOR="#8B0000">**UPDATE**</FONT>: <FONT COLOR="#8B0000">I am finishing my PhD this Fall and will be available for job opportunities starting in December-January. Find my email at the bottom of the page. </FONT> 229 </p> --> 230 231 <!-- Find my CV and contacts <span class="link-highlights"><a href="/cv.html">here</a></span> --> 232 <!-- #8B0000 --> 233 </td> 234 <hr style="border: 0.5px #363636 dashed; margin: 15px 10%;"> 235 </div> 236</div> 237 238<style> 239#typing-effect { 240 display: inline-block; 241 width: 7ch; /* Adjust based on the number of characters */ 242 border-right: 2px solid black; 243 white-space: nowrap; 244 overflow: hidden; 245 animation: typing 1s steps(7, end) forwards, blink-caret 0.75s step-end infinite; 246} 247 248@keyframes typing { 249 from { width: 0; } 250 to { width: 7ch; } /* width as per character count */ 251} 252 253@keyframes blink-caret { 254 from, to { border-color: transparent } 255 50% { border-color: orange } 256} 257 258</style> 259
260<script> 261document.addEventListener("DOMContentLoaded", function() { 262 const element = document.getElementById('typing-effect'); 263 const textWidth = element.scrollWidth; // Get the natural width of the text content 264 265 element.style.width = `${textWidth}px`; // Set the width explicitly to stop at the end of the text 266}); 267// document.addEventListener("DOMContentLoaded", function() { 268// function typeText(element, text, index) { 269// if (index < text.length) { 270// element.innerHTML += text.charAt(index); // Add the next character 271// setTimeout(function() { typeText(element, text, index + 1); }, 100); // Delay before the next character 272// } 273// } 274 275// const header = document.getElementById('typewriter-header'); 276// const text = header.innerText; // Capture the text you want to type 277// header.innerHTML = ''; // Clear the header text 278// typeText(header, text, 0); // Start the typewriter effect 279// }); 280 281document.addEventListener("DOMContentLoaded", function() { 282 function typeElement(typewriterText, content, index, callback) { 283 if (index < content.length) { 284 let element = content[index]; 285 // Check if it's a complete HTML tag 286 if (element.startsWith('<') && element.endsWith('>')) { 287 // It's a complete tag or a text within a tag 288 typewriterText.innerHTML += element; 289 setTimeout(() => typeElement(typewriterText, content, index + 1, callback), 40); 290 } else { 291 // It's regular text 292 typewriterText.innerHTML += element; 293 setTimeout(() => typeElement(typewriterText, content, index + 1, callback), 20); 294 } 295 } else { 296 // All text has been processed, perform any callback if provided 297 if (callback) { 298 callback(); 299 } 300 } 301 } 302 303 304 305 function startTypingParagraph1() { 306 const paragraph1 = document.getElementById('typewriter-paragraph-1'); 307 const originalHeight = paragraph1.clientHeight; 308 paragraph1.style.height = `${originalHeight}px`; 309 paragraph1.innerHTML = ''; // Clear initial content 310 paragraph1.style.opacity = 1; // Ensure visibility 311 const content1 = [ 312 'I', 'â', 'm', ' ', 'N', 'i', 'k', 'i', 't', 'a', ' ', 'D', 'u', 'r', 'a', 's', 'o', 'v', ',', ' ', 'a', ' ', 313 '<strong>PhD student</strong>', ' ', 'a', 't', ' ', '<span class="link-highlights"><a href="https://www.epfl.ch/schools/ic/">EPFL IC</a></span>', 314 ' ', '(', 'ð¨', 'ð', ')', ' ', 'w', 'h', 'e', 'r', 'e', ' ', 'I', ' ', 'w', 'o', 'r', 'k', ' ', 'u', 'n', 'd', 'e', 'r', ' ', 315 't', 'h', 'e', ' ', 's', 'u', 'p', 'e', 'r', 'v', 'i', 's', 'i', 'o', 'n', ' ', 'o', 'f', ' ', 'P', 'r', 'o', 'f', 'e', 's', 's', 'o', 'r', ' ', 316 '<span class="link-highlights"><a href="https://people.epfl.ch/pascal.fua">Pascal Fua</a></span>', ' ', 'a', 't', ' ', 't', 'h', 'e', ' ', 317 '<span class="link-highlights"><a href="https://www.epfl.ch/labs/cvlab/">Computer Vision Laboratory</a></span>', ' ', 'o', 'n', ' ', 'p', 'r', 'o', 'b', 'l', 'e', 'm', 's', ' ', 'r', 'e', 'l', 'a', 't', 'e', 'd', ' ', 't', 'o', ' ', 318 'D', 'e', 'e', 'p', ' ', 'L', 'e', 'a', 'r', 'n', 'i', 'n', 'g', ' ', '(', 'ð¤', ')', ' ', 'a', 'n', 'd', ' ', 'C', 'o', 'm', 'p', 'u', 't', 'e', 'r', ' ', 319 'V', 'i', 's', 'i', 'o', 'n', ' ', '(', 'ð', ')', '.' 320 ]; 321 typeElement(paragraph1, content1, 0, startTypingParagraph2); 322 } 323 324 function startTypingParagraph2() { 325 const paragraph2 = document.getElementById('typewriter-paragraph-2'); 326 const originalHeight = paragraph2.clientHeight; 327 paragraph2.style.height = `${originalHeight}px`; 328 paragraph2.innerHTML = ''; // Clear initial content 329 paragraph2.style.opacity = 1; // Ensure visibility 330 const content2 = [ 331 'P', 'r', 'e', 'v', 'i', 'o', 'u', 's', 'l', 'y', ',', ' ', 'I', ' ', 'w', 'a', 's', ' ', 'a', ' ', 'r', 'e', 's', 'e', 'a', 'r', 'c', 'h', ' ', 's', 'c', 'i', 'e', 'n', 't', 'i', 's', 't', ' ', 'a', 't', ' ', 332 '<span class="link-highlights"><a href="https://machinelearning.apple.com/">Apple</a></span>', ',', ' ', 333 '<span class="link-highlights"><a href="https://www.aboutamazon.com/news/transportation/amazon-prime-air-prepares-for-drone-deliveries">Amazon</a></span>', ',', ' ', 'a', 'n', 'd', ' ', 334 '<span class="link-highlights"><a href="https://news.samsung.com/global/samsung-electronics-launches-ai-center-in-russia">Samsung AI</a></span>', ',', ' ', 'w', 'o', 'r', 'k', 'i', 'n', 'g', ' ', 'o', 'n', ' ', 'p', 'r', 'o', 'b', 'l', 'e', 'm', 's', ' ', 'r', 'e', 'l', 'a', 't', 'e', 'd', ' ', 't', 'o', ' ', 335 'C', 'o', 'm', 'p', 'u', 't', 'e', 'r', ' ', 'V', 'i', 's', 'i', 'o', 'n', ' ', 'a', 'n', 'd', ' ', 'D', 'e', 'e', 'p', ' ', 'L', 'e', 'a', 'r', 'n', 'i', 'n', 'g', '.', ' ', 'I', ' ', 'c', 'o', 'm', 'p', 'l', 'e', 't', 'e', 'd', ' ', 'm', 'y', ' ', 'b', 'a', 'c', 'h', 'e', 'l', 'o', 'r', '\'', 's', ' ', 'd', 'e', 'g', 'r', 'e', 'e', ' ', 'a', 't', ' ', '<span class="link-highlights"><a href="https://eng.mipt.ru/">MIPT</a></span>', ' ', '(', 'ð·', 'ðº', ')', ' ', 'i', 'n', ' ', 'A', 'p', 'p', 'l', 'i', 'e', 'd', ' ', 'P', 'h', 'y', 's', 'i', 'c', 's', ' ', 'a', 'n', 'd', ' ', 'M', 'a', 't', 'h', 'e', 'm', 'a', 't', 'i', 'c', 's', ' ', 'w', 'h', 'i', 'l', 'e', ' ', 'a', 'l', 's', 'o', ' ', 's', 't', 'u', 'd', 'y', 'i', 'n', 'g', ' ', 'a', 't', ' ', 't', 'h', 'e', ' ', '<span class="link-highlights"><a href="https://yandexdataschool.com/">Yandex School of Data Analysis</a></span>', '.' 336 ]; 337 typeElement(paragraph2, content2, 0, startTypingParagraph3); 338 } 339 340 function startTypingParagraph3() { 341 const paragraph3 = document.getElementById('typewriter-paragraph-3'); 342 const originalHeight = paragraph3.clientHeight; 343 paragraph3.style.height = `${originalHeight}px`; 344 paragraph3.innerHTML = ''; // Clear initial content 345 paragraph3.style.opacity = 1; // Ensure visibility 346 const content3 = [ 347 '<FONT COLOR="#8B0000">*UPDATE*</FONT>', ':', ' ', 'U', 'n', 't', 'i', 'l', ' ', 'n', 'e', 'x', 't', ' ', 'f', 'a', 'l', 'l', ',', ' ', 'I', '\'', 'l', 'l', ' ', 'b', 'e', ' ', 'w', 'i', 't', 'h', ' ', '<span class="link-highlights">
347<a href="https://www.nvidia.com/en-us/self-driving-cars/">Nvidia</a></span>', '\'', 's', ' ', 'a', 'u', 't', 'o', 'n', 'o', 'm', 'o', 'u', 's', ' ', 'd', 'r', 'i', 'v', 'i', 'n', 'g', ' ', 'd', 'e', 'p', 'a', 'r', 't', 'm', 'e', 'n', 't', ',', ' ', 'f', 'o', 'c', 'u', 's', 'i', 'n', 'g', ' ', 'o', 'n', ' ', 'i', 'm', 'p', 'r', 'o', 'v', 'i', 'n', 'g', ' ', 'p', 'e', 'r', 'c', 'e', 'p', 't', 'i', 'o', 'n', ' ', 'm', 'o', 'd', 'e', 'l', 's', '.' 348 ]; 349 typeElement(paragraph3, content3, 0, null); // No callback, this is the last one 350 } 351 352 // Start the first animation 353 // startTypingParagraph1(); 354}); 355 356// document.addEventListener("DOMContentLoaded", function() { 357// function typeElement(typewriterText, content, index) { 358// if (index < content.length) { 359// let element = content[index]; 360// if (element.startsWith('<') && element.endsWith('>')) { 361// // It's a complete tag or a text within a tag 362// typewriterText.innerHTML += element; 363// setTimeout(function() { typeElement(typewriterText, content, index + 1); }, 10); 364// } else { 365// // Regular text outside tags 366// typewriterText.innerHTML += element; 367// setTimeout(function() { typeElement(typewriterText, content, index + 1); }, 3); 368// } 369// } 370// } 371 372// const paragraph1 = document.getElementById('typewriter-paragraph-1'); 373// paragraph1.innerHTML = ''; // Clear initial content 374// paragraph1.style.opacity = 1; // Ensure visibility 375// const content1 = [ 376// 'I', 'â', 'm', ' ', 'N', 'i', 'k', 'i', 't', 'a', ' ', 'D', 'u', 'r', 'a', 's', 'o', 'v', ',', ' ', 'a', ' ', 377// '<strong>PhD student</strong>', ' ', 'a', 't', ' ', '<span class="link-highlights"><a href="https://www.epfl.ch/schools/ic/">EPFL IC</a></span>', 378// ' ', '(', 'ð¨', 'ð', ')', ' ', 'w', 'h', 'e', 'r', 'e', ' ', 'I', ' ', 'w', 'o', 'r', 'k', ' ', 'u', 'n', 'd', 'e', 'r', ' ', 379// 't', 'h', 'e', ' ', 's', 'u', 'p', 'e', 'r', 'v', 'i', 's', 'i', 'o', 'n', ' ', 'o', 'f', ' ', 'P', 'r', 'o', 'f', 'e', 's', 's', 'o', 'r', ' ', 380// '<span class="link-highlights"><a href="https://people.epfl.ch/pascal.fua">Pascal Fua</a></span>', ' ', 'a', 't', ' ', 't', 'h', 'e', ' ', 381// '<span class="link-highlights"><a href="https://www.epfl.ch/labs/cvlab/">Computer Vision Laboratory</a></span>', ' ', 'o', 'n', ' ', 'p', 'r', 'o', 'b', 'l', 'e', 'm', 's', ' ', 'r', 'e', 'l', 'a', 't', 'e', 'd', ' ', 't', 'o', ' ', 382// 'D', 'e', 'e', 'p', ' ', 'L', 'e', 'a', 'r', 'n', 'i', 'n', 'g', ' ', '(', 'ð¤', ')', ' ', 'a', 'n', 'd', ' ', 'C', 'o', 'm', 'p', 'u', 't', 'e', 'r', ' ', 383// 'V', 'i', 's', 'i', 'o', 'n', ' ', '(', 'ð', ')', '.' 384// ]; 385// setTimeout(function() { typeElement(paragraph1, content1, 0); }, 20); 386 387// const paragraph2 = document.getElementById('typewriter-paragraph-2'); 388// paragraph2.innerHTML = ''; // Clear initial content 389// paragraph2.style.opacity = 1; // Ensure visibility 390// const content2 = [ 391// 'P', 'r', 'e', 'v', 'i', 'o', 'u', 's', 'l', 'y', ',', ' ', 'I', ' ', 'w', 'a', 's', ' ', 'a', ' ', 'r', 'e', 's', 'e', 'a', 'r', 'c', 'h', ' ', 's', 'c', 'i', 'e', 'n', 't', 'i', 's', 't', ' ', 'a', 't', ' ', 392// '<span class="link-highlights"><a href="https://machinelearning.apple.com/">Apple</a></span>', ',', ' ', 393// '<span class="link-highlights"><a href="https://www.aboutamazon.com/news/transportation/amazon-prime-air-prepares-for-drone-deliveries">Amazon</a></span>', ',', ' ', 'a', 'n', 'd', ' ', 394// '<span class="link-highlights"><a href="https://news.samsung.com/global/samsung-electronics-launches-ai-center-in-russia">Samsung AI</a></span>', ',', ' ', 'w', 'o', 'r', 'k', 'i', 'n', 'g', ' ', 'o', 'n', ' ', 'p', 'r', 'o', 'b', 'l', 'e', 'm', 's', ' ', 'r', 'e', 'l', 'a', 't', 'e', 'd', ' ', 't', 'o', ' ', 395// 'C', 'o', 'm', 'p', 'u', 't', 'e', 'r', ' ', 'V', 'i', 's', 'i', 'o', 'n', ' ', 'a', 'n', 'd', ' ', 'D', 'e', 'e', 'p', ' ', 'L', 'e', 'a', 'r', 'n', 'i', 'n', 'g', '.', ' ', 'I', ' ', 'c', 'o', 'm', 'p', 'l', 'e', 't', 'e', 'd', ' ', 'm', 'y', ' ', 'b', 'a', 'c', 'h', 'e', 'l', 'o', 'r', '\'', 's', ' ', 'd', 'e', 'g', 'r', 'e', 'e', ' ', 'a', 't', ' ', '<span class="link-highlights"><a href="https://eng.mipt.ru/">MIPT</a></span>', ' ', '(', 'ð·', 'ðº', ')', ' ', 'i', 'n', ' ', 'A', 'p', 'p', 'l', 'i', 'e', 'd', ' ', 'P', 'h', 'y', 's', 'i', 'c', 's', ' ', 'a', 'n', 'd', ' ', 'M', 'a', 't', 'h', 'e', 'm', 'a', 't', 'i', 'c', 's', ' ', 'w', 'h', 'i', 'l', 'e', ' ', 'a', 'l', 's', 'o', ' ', 's', 't', 'u', 'd', 'y', 'i', 'n', 'g', ' ', 'a', 't', ' ', 't', 'h', 'e', ' ', '<span class="link-highlights"><a href="https://yandexdataschool.com/">Yandex School of Data Analysis</a></span>', '.' 396// ]; 397// setTimeout(function() { typeElement(paragraph2, content2, 0); }, 20); 398 399// const paragraph3 = document.getElementById('typewriter-paragraph-3'); 400// paragraph3.innerHTML = ''; // Clear initial content 401// paragraph3.style.opacity = 1; // Ensure visibility 402// const content3 = [ 403// '<FONT COLOR="#8B0000">*UPDATE*</FONT>', ':', ' ', 'U', 'n', 't', 'i', 'l', ' ', 'n', 'e', 'x', 't', ' ', 'f', 'a', 'l', 'l', ',', ' ', 'I', '\'', 'l', 'l', ' ', 'b', 'e', ' ', 'w', 'i', 't', 'h', ' ', '<span class="link-highlights">
403<a href="https://www.nvidia.com/en-us/self-driving-cars/">Nvidia</a></span>', '\'', 's', ' ', 'a', 'u', 't', 'o', 'n', 'o', 'm', 'o', 'u', 's', ' ', 'd', 'r', 'i', 'v', 'i', 'n', 'g', ' ', 'd', 'e', 'p', 'a', 'r', 't', 'm', 'e', 'n', 't', ',', ' ', 'f', 'o', 'c', 'u', 's', 'i', 'n', 'g', ' ', 'o', 'n', ' ', 'i', 'm', 'p', 'r', 'o', 'v', 'i', 'n', 'g', ' ', 'p', 'e', 'r', 'c', 'e', 'p', 't', 'i', 'o', 'n', ' ', 'm', 'o', 'd', 'e', 'l', 's', '.' 404// ]; 405// setTimeout(function() { typeElement(paragraph3, content3, 0); }, 20); 406// }); 407</script>
407 408 409<!-- <div id="carouselContainer" class="carousel-container"> 410 <div class="carousel-row"> 411 <div class="col-4 align-self-center"> 412 <img style="max-height: 100px; width: auto;" alt="NVIDIA logo" src="/assets/images/nvidia_logo.png"> 413 </div> 414 <div class="col-4 align-self-center"> 415 <img style="max-height: 100px; width: auto;" alt="Amazon logo" src="/assets/images/Amazon_logo.svg.png"> 416 </div> 417 <div class="col-4 align-self-center" style="padding-left: 7%;"> 418 <img style="max-height: 100px; width: auto;" alt="EPFL logo" src="/assets/images/364px-Logo_EPFL.svg.png"> 419 </div> 420 <div class="col-4 align-self-center"> 421 <img style="max-height: 100px; width: auto;" alt="Apple logo" src="/assets/images/Apple_logo_black.svg.png"> 422 </div> 423 <div class="col-4 align-self-center"> 424 <img style="max-height: 100px; width: auto;" alt="Samsung logo" src="/assets/images/Samsung_Logo.svg.png"> 425 </div> 426 <div class="col-4 align-self-center"> 427 <img style="max-height: 100px; width: auto;" alt="NVIDIA logo" src="/assets/images/nvidia_logo.png"> 428 </div> 429 <div class="col-4 align-self-center"> 430 <img style="max-height: 100px; width: auto;" alt="Amazon logo" src="/assets/images/Amazon_logo.svg.png"> 431 </div> 432 <div class="col-4 align-self-center" style="padding-left: 7%;"> 433 <img style="max-height: 100px; width: auto;" alt="EPFL logo" src="/assets/images/364px-Logo_EPFL.svg.png"> 434 </div> 435 <div class="col-4 align-self-center"> 436 <img style="max-height: 100px; width: auto;" alt="Apple logo" src="/assets/images/Apple_logo_black.svg.png"> 437 </div> 438 <div class="col-4 align-self-center"> 439 <img style="max-height: 100px; width: auto;" alt="Samsung logo" src="/assets/images/Samsung_Logo.svg.png"> 440 </div> 441 </div> 442</div> --> 443<br> 444<div id="carouselContainer" class="carousel-container"> 445 <div class="carousel-row"> 446 <div class="col-4 align-self-center"> 447 <img style="width: 100px;" alt="NVIDIA logo" src="/assets/images/nvidia_logo.png"> 448 </div> 449 <div class="col-4 align-self-center"> 450 <img style="width: 150px" alt="Amazon logo" src="/assets/images/Amazon_logo.svg.png"> 451 </div> 452 <div class="col-4 align-self-center" style="padding-left: 7%;"> 453 <img style="width: 150px" alt="EPFL logo" src="/assets/images/364px-Logo_EPFL.svg.png"> 454 </div> 455 <div class="col-4 align-self-center"> 456 <img style="width: 70px" alt="Apple logo" src="/assets/images/Apple_logo_black.svg.png"> 457 </div> 458 <div class="col-4 align-self-center"> 459 <img style="width: 150px" alt="Samsung logo" src="/assets/images/Samsung_Logo.svg.png"> 460 </div> 461 <div class="col-4 align-self-center"> 462 <img style="width: 150px" alt="Yandex logo" src="/assets/images/Yandex_logo.svg.png"> 463 </div> 464 <div class="col-4 align-self-center"> 465 <img style="width: 100px" alt="NVIDIA logo" src="/assets/images/nvidia_logo.png"> 466 </div> 467 <div class="col-4 align-self-center"> 468 <img style="width: 150px" alt="Amazon logo" src="/assets/images/Amazon_logo.svg.png"> 469 </div> 470 <div class="col-4 align-self-center" style="padding-left: 7%;"> 471 <img style="width: 150px" alt="EPFL logo" src="/assets/images/364px-Logo_EPFL.svg.png"> 472 </div> 473 <div class="col-4 align-self-center"> 474 <img style="width: 70px" alt="Apple logo" src="/assets/images/Apple_logo_black.svg.png"> 475 </div> 476 <div class="col-4 align-self-center"> 477 <img style="width: 150px" alt="Samsung logo" src="/assets/images/Samsung_Logo.svg.png"> 478 </div> 479 <div class="col-4 align-self-center"> 480 <img style="width: 150px" alt="Yandex logo" src="/assets/images/Yandex_logo.svg.png"> 481 </div> 482 </div> 483</div> 484 485<!-- DETECTION VISUALIZATION --> 486<!-- <div class="container-fluid" style="width: 100%;"> 487 <div class="row" style="width: 100%; height: 100%;"> 488 <div class="col-sm-12 align-self-center"> 489 <div id="container" style="width: 100%; height: 300px;"></div> 490 <div class="col-sm-12 text-center" style="padding-top: 5px; width: 100%;"> 491 <div> 492 <span style="display: inline-block; width: 12px; height: 12px; background-color: #04d904; border-radius: 50%; margin-right: 5px;"></span>Detected objects 493 </div> 494 </div> 495 </div> 496 </div> 497</div>
498<script src="https://cdnjs.cloudflare.com/ajax/libs/three.js/r128/three.min.js"></script>
vendor: 1 bytes, line 498
498
499<script src="https://cdn.rawgit.com/mrdoob/three.js/r128/examples/js/loaders/OBJLoader.js"></script>
vendor: 1 bytes, line 499
499
500<script src="https://cdn.rawgit.com/mrdoob/three.js/r128/examples/js/controls/OrbitControls.js"></script>
vendor: 1 bytes, line 500
500
501<script src="https://cdn.jsdelivr.net/gh/mrdoob/three.js@r128/examples/js/renderers/CSS2DRenderer.js"></script>
vendor: 1 bytes, line 501
501
502<script type="module"> 503 import { CSS2DRenderer, CSS2DObject } from 'https://cdn.jsdelivr.net/gh/mrdoob/three.js@r128/examples/jsm/renderers/CSS2DRenderer.js'; 504 // Scene, Camera, Renderer 505 const scene = new THREE.Scene(); 506 const camera = new THREE.PerspectiveCamera(75, window.innerWidth / window.innerHeight, 0.1, 1000); 507 const renderer = new THREE.WebGLRenderer({ antialias: true, alpha: true }); 508 const container = document.getElementById('container'); 509 renderer.setSize(container.clientWidth, container.clientHeight); 510 renderer.setClearColor(0x000000, 0); // Set clear color to transparent 511 container.appendChild(renderer.domElement); 512 513 // Orbit Controls 514 const controls = new THREE.OrbitControls(camera, renderer.domElement); 515 controls.enableDamping = true; 516 controls.dampingFactor = 0.25; 517 controls.enableZoom = true; 518 controls.minPolarAngle = Math.PI / 2 - 0.1; // Restrict vertical rotation 519 controls.maxPolarAngle = Math.PI / 2 + 0.1; // Restrict vertical rotation 520 521 const axesHelper = new THREE.AxesHelper(5); 522 scene.add(axesHelper); 523 524 // Function to add labels to the axes 525 function createAxisLabel(text, position) { 526 const loader = new THREE.FontLoader(); 527 loader.load('https://threejs.org/examples/fonts/helvetiker_regular.typeface.json', function (font) { 528 const textGeometry = new THREE.TextGeometry(text, { 529 font: font, 530 size: 0.2, 531 height: 0.05, 532 }); 533 const textMaterial = new THREE.MeshBasicMaterial({ color: 0x000000 }); 534 const textMesh = new THREE.Mesh(textGeometry, textMaterial); 535 textMesh.position.copy(position); 536 scene.add(textMesh); 537 }); 538 } 539 540 // Add labels to axes 541 createAxisLabel('X', new THREE.Vector3(5, 0, 0)); 542 createAxisLabel('Y', new THREE.Vector3(0, 5, 0)); 543 createAxisLabel('Z', new THREE.Vector3(0, 0, 5)); 544 545 546 // Load .obj file 547 const objLoader = new THREE.OBJLoader(); 548 objLoader.load('/assets/geometries/test_points.obj', (object) => { 549 object.scale.set(1, 1, 1); // Adjust the scale if necessary 550 object.rotation.x = -Math.PI / 2; 551 object.position.set(0, 0, 0); // Adjust the position if necessary 552 553 scene.add(object); 554 console.log('OBJ file loaded successfully'); 555 }, (xhr) => { 556 console.log((xhr.loaded / xhr.total * 100) + '% loaded'); 557 }, (error) => { 558 console.error('An error happened', error); 559 }); 560 561 // Load second .obj file 562 const objLoader2 = new THREE.OBJLoader(); 563 objLoader2.load('/assets/geometries/test_100_gt.obj', (object2) => { 564 object2.scale.set(1, 1, 1); // Adjust the scale if necessary 565 // object2.rotation.x = -Math.PI / 2; 566 object2.position.set(0, 0, 0); // Adjust the position if necessary 567 // scene.add(object2); 568 569 // Edges Helper for second object 570 object2.traverse((child) => { 571 if (child.isMesh) { 572 const geometry = child.geometry.clone(); 573 const position = geometry.attributes.position; 574 575 for (let i = 0; i < position.count; i++) { 576 const x = position.getX(i); 577 const y = position.getY(i); 578 const z = position.getZ(i); 579 580 position.setXYZ(i, x, z, -y); // Swap y and z 581 } 582 583 const edges = new THREE.EdgesGeometry(geometry); 584 const lineMaterial = new THREE.LineBasicMaterial({ color: 0x00ff00 }); 585 const lines = new THREE.LineSegments(edges, lineMaterial); 586 587 lines.position.copy(child.position); 588 lines.rotation.copy(child.rotation); 589 lines.scale.copy(child.scale); 590 591 scene.add(lines); 592 } 593 }); 594 595 console.log('Second OBJ file loaded successfully'); 596 }, (xhr) => { 597 console.log((xhr.loaded / xhr.total * 100) + '% loaded'); 598 }, (error) => { 599 console.error('An error happened', error); 600 }); 601 602 // Set the initial polar angle 603 camera.position.set(0, 10, 25); // Adjust the position as needed 604 controls.target.set(0, 10, 0); // Adjust the target as needed 605 controls.update(); 606 607 // Lighting 608 const ambientLight = new THREE.AmbientLight(0x404040, 1); // Ambient light for general illumination 609 scene.add(ambientLight); 610 611 const directionalLight = new THREE.DirectionalLight(0xffffff, 1); // Directional light to highlight the object 612 directionalLight.position.set(10, 10, 10).normalize(); 613 scene.add(directionalLight); 614 615 // Render Loop 616 const animate = function () { 617 requestAnimationFrame(animate); 618 controls.update(); 619 renderer.render(scene, camera); 620 }; 621 622 animate(); 623 624 // Handle window resize 625 window.addEventListener('resize', () => { 626 camera.aspect = container.clientWidth / container.clientHeight; 627 camera.updateProjectionMatrix(); 628 renderer.setSize(container.clientWidth, container.clientHeight); 629 }); 630 631</script>
631 --> 632 633 634 635 636<style> 637.carousel-container { 638 overflow: hidden; 639 white-space: nowrap; 640 position: relative; 641 background: linear-gradient(to right, #69b0fc24, #331a6519); 642 padding-top: 5px; 643 padding-bottom: 5px; 644 border-radius: 5px; 645} 646 647.carousel-row { 648 display: flex; 649 animation: scroll 8s linear infinite; 650 /* background-color: #b2c7dd0c; */ 651} 652 653@keyframes scroll { 654 0% { 655 transform: translateX(0); 656 } 657 100% { 658 transform: translateX(-210%); /* Reset to the starting position smoothly */ 659 } 660} 661 662/* .col-4 { 663 display: inline-block; 664 width: 10%; 665 background-color: #f8f9fa71; 666 padding: 10px; 667 border-radius: 5px; 668} */ 669 670.col-4 { 671 display: flex; 672 justify-content: center; 673 align-items: center; 674 675 background-color: #b2c7dd00; 676 padding: 5px; 677 margin: 5px; 678 box-sizing: border-box; 679 border-radius: 5px; 680 /* box-shadow: 0 2px 5px rgba(0,0,0,0.1); */ 681} 682 683</style> 684 685 686<h2 class="divider line glow" contenteditable>Research</h2> 687 688 689<hr style="border: 0.01px #363636 dashed; margin: 15px 10%;"> 690 691<div class="row"> 692 <div class="col-12 align-self-center"> 693 <p style="text-align: left;"> 694 Iâm interested in advancing <strong>vision-language models (VLMs)</strong> and <strong>video understanding</strong>, building scalable retrieval and reasoning pipelines that turn complex video data into <strong>searchable, interpretable knowledge</strong>. 695 </p> 696 697 <p style="text-align:left;"> 698 699 My work explores <strong>robustness</strong>, <strong>generalization</strong>, and <strong>semantic search</strong> to make multimodal AI systems more powerful and reliable. 700 701 For a non-exhaustive list of my research interests, refer to my <a href="https://scholar.google.com/citations?user=KMMvTfcAAAAJ&hl=en">Google Scholar</a>. 702 </p> 703 704 705 <!-- <p style="text-align:left;"> 706 I'm broadly interested in enhancing the <span class="topic-highlights">generalization</span> and <span class="topic-highlights">robustness</span> of modern deep learning models through <span class="topic-highlights">uncertainty quantification</span> and <span class="topic-highlights">bayesian methods</span>. 707 </p> 708 <p style="text-align:left;"> 709 Consequently, I'm also curious about new approaches to <span class="topic-highlights">active learning</span>, <span class="topic-highlights">bayesian optimization</span>, and other <span class="topic-highlights">data-efficient</span> techniques for various domains. 710 </p> 711 <p style="text-align:left;"> 712 Mostly, my research is focused on <span class="topic-highlights">computer vision</span> and <span class="topic-highlights">image processing</span> tasks, but it is not limited to them. Occasionally, I delve into <span class="topic-highlights">natural language processing</span> and <span class="topic-highlights">reinforcement learning</span>. 713 </p> --> 714 <p style="text-align:left;"> 715 I have also contributed as a reviewer for various academic conferences and journals, such as: <span class="topic-highlights">CVPR</span> <span class="topic-highlights">NeurIPS</span> <span class="topic-highlights">ICLR</span> <span class="topic-highlights">ICML</span> <span class="topic-highlights">ICCV</span> <span class="topic-highlights">ECCV</span> <span class="topic-highlights">ICRA</span> <span class="topic-highlights">IROS</span> <span class="topic-highlights">ACCV</span> <span class="topic-highlights">RA-L</span> <span class="topic-highlights">IJCV</span> <span class="topic-highlights">AISTATS</span> <span class="topic-highlights">TPAMI</span> <span class="topic-highlights">TMLR</span> <span class="topic-highlights">WACV</span>, and other (<a href="/review_list.html">full list</a>). 716 </p> 717 718 719 <div class="row"> 720 <div class="col-12 align-self-center"> 721 <p style="text-align:center; margin-top:10px"><u><span class="topic-highlights">My Stats from 722 <a href="https://scholar.google.com/citations?user=KMMvTfcAAAAJ&hl=en">Google Scholar</a>:</span></u></p> 723 <!-- <h3 style="text-align:center; margin-bottom:20px;">My Stats from Google Scholar:</h3> --> 724 </div> 725 <div class="col-sm-4 align-self-center text-center"> 726 <ul class="list-inline d-sm-inline-block d-block my-stats-list"> 727 <li class="list-inline-item"><span class="topic-highlights">Citations: -</span></li> 728 <li class="list-inline-item"><span class="topic-highlights">h-index: -</li> 729 <li class="list-inline-item"><span class="topic-highlights">i10-index: -</span></li> 730 <!-- Add more stats as needed --> 731 </ul> 732 </div> 733 <div class="col-sm-8 align-self-center">
734 <canvas id="citationChart" width="300" height="120"></canvas> 735 </div> 736 </div> 737 738 </div> 739</div> 740 741<hr style="border: 0.01px #363636 dashed; margin: 0px 10%;"> 742 743<h2 class="divider line glow" contenteditable>Publications</h2> 744 745<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 746 <div class="col-sm-4 align-self-center"> 747 <center> 748 <img style="width:100%; padding: 10%; border-radius: 65%" 749 alt="depth estimation photo" src="/assets/images/bayesian_deep_learning_book_round.jpeg"> 750 </center> 751 </div> 752 753 <div class="col-sm-8 align-self-center"> 754 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Handbook of Bayesian Deep Learning</span></b></h6> 755 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> Kevin Murphy, Yarin Gal, Andrew Gordon Wilson, Philipp Hennig, Vincent Fortuin, <b style="color: white">Nikita Durasov</b>, Theodore Papamarkou, et al. 756 <br> 757 </p> 758 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">BOOK 2026</span></p> 759 <p style="text-align: center;"> 760 <a href="https://zenodo.org/records/22114549?preview_file=handbook_of_bdl.pdf"> pdf </a> 761 | 762 <a href="https://zenodo.org/records/22114549"> zenodo </a> 763 | 764 <a href="https://bdlbook.org/"> web </a> 765 <!-- 766 | 767 <a href="https://arxiv.org/pdf/2403.16732.pdf"> pdf </a> 768 | 769 <a href="https://www.neuralconcept.com/post/neural-concept-presents-enabling-uncertainty-estimation-in-iterative-neural-networks-technical-paper-at-international-conference-on-machine-learning"> media </a> 770 | 771 <a href="https://www.youtube.com/watch?v=27i0IBh65Pw&t=3s"> video </a> 772 | 773 <a href="/projects/ittt/"> page </a> 774 --> 775 </p> 776 </div> 777</div> 778 779 780<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 781 <div class="col-sm-4 align-self-center"> 782 <center> 783 <img style="width:100%; padding: 10%; border-radius: 65%" 784 alt="depth estimation photo" src="/assets/images/diff_relight.png"> 785 </center> 786 </div> 787 788 <div class="col-sm-8 align-self-center"> 789 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Diffusion Models are Secretly Zero-Shot 3DGS Harmonizers</span></b></h6> 790 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> Vsevolod Skorokhodov<sup>*</sup>, <b style="color: white">Nikita Durasov<sup>*</sup></b>, Pascal Fua 791 <br> 792 </p> 793 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">TMLR 2026</span></p> 794 <p style="text-align: center;"> 795 <a href="https://arxiv.org/abs/2503.06740"> pdf </a> 796 | 797 <a href="https://github.com/sevashasla/D3DR"> code </a> 798 | 799 <a href="/projects/diff_relight/"> page </a> 800 <!-- 801 | 802 <a href="https://arxiv.org/pdf/2403.16732.pdf"> pdf </a> 803 | 804 <a href="https://www.neuralconcept.com/post/neural-concept-presents-enabling-uncertainty-estimation-in-iterative-neural-networks-technical-paper-at-international-conference-on-machine-learning"> media </a> 805 | 806 <a href="https://www.youtube.com/watch?v=27i0IBh65Pw&t=3s"> video </a> 807 | 808 <a href="/projects/ittt/"> page </a> 809 --> 810 </p> 811 </div> 812</div> 813 814<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 815 <div class="col-sm-4 align-self-center"> 816 <center> 817 <img style="width:100%; padding: 10%; border-radius: 65%" 818 alt="depth estimation photo" src="/assets/images/unc_det_car.png"> 819 </center> 820 </div> 821 <div class="col-sm-8 align-self-center"> 822 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Uncertainty-Guided Data Curation for 3D Object Detection</span></b></h6> 823 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 824 <b style="color: white">Nikita Durasov</b>, Rafid Mahmood, Jiwoong Choi, Marc T Law, James Lucas, Pascal Fua, Jose M Alvarez 825 <br> 826 </p> 827 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">CVPRW 2026</span></p> 828 <p style="text-align: center;"> 829 <a href="https://arxiv.org/pdf/2410.23910"> pdf </a> 830 | 831 <a href="https://patents.google.com/patent/US20260111771A1/en"> patent </a> 832 | 833 <a href="/projects/unc3d/"> page </a> 834 <!-- 835 <a href="https://github.com/nikitadurasov/ittt">
835 code </a> 836 | 837 <a href="https://arxiv.org/pdf/2403.16732.pdf"> pdf </a> 838 | 839 <a href="https://www.neuralconcept.com/post/neural-concept-presents-enabling-uncertainty-estimation-in-iterative-neural-networks-technical-paper-at-international-conference-on-machine-learning"> media </a> 840 | 841 <a href="https://www.youtube.com/watch?v=27i0IBh65Pw&t=3s"> video </a> 842 | 843 <a href="/projects/ittt/"> page </a> 844 --> 845 </p> 846 </div> 847</div> 848 849<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 850 <div class="col-sm-4 align-self-center"> 851 <center> 852 <img style="width:100%; padding: 10%; border-radius: 65%" 853 alt="depth estimation photo" src="/assets/images/av_captioning.png"> 854 </center> 855 </div> 856 <div class="col-sm-8 align-self-center"> 857 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Scalable Parallel Prompting for Complex AV Video Captioning</span></b></h6> 858 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 859 April Yang<sup>*</sup>, Roberto Amoroso<sup>*</sup>, <b style="color: white">Nikita Durasov</b>, Devansh Bisla, Sandipan Kundu, Elmar Haussmann, Ruchi Bhargava, Maying Shen, Nadine Chang, Jose M. Alvarez 860 <br> 861 </p> 862 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">CVPRW 2026</span></p> 863 <p style="text-align: center;"> 864 <a href="https://openaccess.thecvf.com/content/CVPR2026W/DataMFM/papers/Yang_Scalable_Parallel_Prompting_for_Complex_AV_Video_Captioning_CVPRW_2026_paper.pdf"> pdf </a> 865 <!-- | 866 <a href="https://patents.google.com/patent/US20260111771A1/en"> patent </a> 867 | 868 <a href="/projects/unc3d/"> page </a> --> 869 <!-- 870 <a href="https://github.com/nikitadurasov/ittt"> code </a> 871 | 872 <a href="https://arxiv.org/pdf/2403.16732.pdf"> pdf </a> 873 | 874 <a href="https://www.neuralconcept.com/post/neural-concept-presents-enabling-uncertainty-estimation-in-iterative-neural-networks-technical-paper-at-international-conference-on-machine-learning"> media </a> 875 | 876 <a href="https://www.youtube.com/watch?v=27i0IBh65Pw&t=3s"> video </a> 877 | 878 <a href="/projects/ittt/"> page </a> 879 --> 880 </p> 881 </div> 882</div> 883 884<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 885 <div class="col-sm-4 align-self-center"> 886 <center> 887 <img style="width:100%; padding: 10%; border-radius: 65%" 888 alt="depth estimation photo" src="/assets/images/mirrorcheck.png"> 889 </center> 890 </div> 891 <div class="col-sm-8 align-self-center"> 892 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">MirrorCheck: Efficient Adversarial Defense for Vision-Language Models</span></b></h6> 893 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 894 Samar Fares<sup>*</sup>, Klea Ziu<sup>*</sup>, Toluwani Aremu<sup>*</sup>, <b style="color: white">Nikita Durasov</b>, Martin TakáÄ, Pascal Fua, Karthik Nandakumar, Ivan Laptev 895 <br> 896 </p> 897 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">CVPRW 2026 (Paper Award)</span></p> 898 <p style="text-align: center;"> 899 <a href="https://arxiv.org/abs/2406.09250"> pdf </a> 900 | 901 <a class="disabled" href=""> code </a> 902 <!-- | 903 - 904 <a href="https://www.youtube.com/watch?v=Bsi0RMTdEaI"> video </a> --> 905 | 906 <a href="/projects/mirrorcheck/"> page </a> 907 </p> 908 </div> 909</div> 910 911<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 912 <div class="col-sm-4 align-self-center"> 913 <center> 914 <img style="width:100%; padding: 10%; border-radius: 65%" 915 alt="depth estimation photo" src="/assets/images/ittt_logo_snake.jpeg"> 916 </center> 917 </div> 918 <div class="col-sm-8 align-self-center"> 919 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">IT<sup>3</sup>: Idempotent Test-Time Training</span></b></h6> 920 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 921 <b style="color: white">Nikita Durasov<sup>*</sup></b>, Assaf Shocher<sup>*</sup>, Doruk Oner, Gal Chechik, Alexei A. Efros, Pascal Fua 922 <br> 923 </p> 924 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">ICML 2025</span></p> 925 <p style="text-align: center;"> 926 <a href="https://arxiv.org/pdf/2410.04201"> pdf </a> 927 | 928 <a href="https://github.com/nikitadurasov/ittt">
928 code </a> 929 | 930 <!-- 931 <a href="https://www.neuralconcept.com/post/neural-concept-presents-enabling-uncertainty-estimation-in-iterative-neural-networks-technical-paper-at-international-conference-on-machine-learning"> media </a> 932 |--> 933 <a href="https://www.youtube.com/watch?v=eKGKpN8fFRM"> video </a> 934 | 935 <a href="/projects/ittt/"> page </a> 936 </p> 937 </div> 938</div> 939 940<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 941 <div class="col-sm-4 align-self-center"> 942 <center> 943 <img style="width:100%; padding: 10%; border-radius: 65%" 944 alt="depth estimation photo" src="/assets/images/iterative_model.jpg"> 945 </center> 946 </div> 947 <div class="col-sm-8 align-self-center"> 948 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Enabling Uncertainty Estimation in Iterative Neural Networks</span></b></h6> 949 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 950 <b style="color: white">Nikita Durasov</b>, Doruk Oner, Jonathan Donier, Hieu Le, Pascal Fua 951 <br> 952 </p> 953 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">ICML 2024</span></p> 954 <p style="text-align: center;"> 955 <a href="https://arxiv.org/pdf/2403.16732.pdf"> pdf </a> 956 | 957 <a href="https://github.com/cvlab-epfl/iter_unc"> code </a> 958 | 959 <a href="https://www.neuralconcept.com/post/neural-concept-presents-enabling-uncertainty-estimation-in-iterative-neural-networks-technical-paper-at-international-conference-on-machine-learning"> media </a> 960 | 961 <a href="https://www.youtube.com/watch?v=27i0IBh65Pw&t=3s"> video </a> 962 | 963 <a href="/projects/unc_iter/"> page </a> 964 </p> 965 </div> 966</div> 967 968<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 969 <div class="col-sm-4 align-self-center"> 970 <center> 971 <img style="width:100%; padding: 10%; border-radius: 65%" 972 alt="depth estimation photo" src="/assets/images/zigzag_copy.png"> 973 </center> 974 </div> 975 <div class="col-sm-8 align-self-center"> 976 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference</span></b></h6> 977 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 978 <b style="color: white">Nikita Durasov</b>, Nik Dorndorf, Hieu Le, Pascal Fua 979 <br> 980 </p> 981 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">TMLR 2024</span> <span class="topic-highlights">ICMLW 2024</span></p> 982 <p style="text-align: center;"> 983 <a href="https://openreview.net/pdf?id=QSvb6jBXML"> openreview </a> 984 | 985 <a href="https://arxiv.org/pdf/2211.11435.pdf"> pdf </a> 986 | 987 <a href="https://github.com/cvlab-epfl/zigzag"> code </a> 988 | 989 <a href="/projects/zigzag/"> page </a> 990 </p> 991 </div> 992</div> 993 994<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 995 <div class="col-sm-4 align-self-center"> 996 <center> 997 <img style="width:100%; padding: 10%; border-radius: 65%" 998 alt="depth estimation photo" src="/assets/images/facerec.png"> 999 </center> 1000 </div> 1001 <div class="col-sm-8 align-self-center"> 1002 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">How to Boost Face Recognition with StyleGAN?</span></b></h6> 1003 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1004 Artem Sevastopolsky, Yuri Malkov, <b style="color: white">Nikita Durasov</b>, Luisa Verdoliva, Matthias NieÃner 1005 <!-- Sergei Ivanov, <b style="color: white">Nikita Durasov</b>, Evgeny Burnaev --> 1006 <br> 1007 </p> 1008 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">ICCV 2023</span></p> 1009 <p style="text-align: center;"> 1010 <a href="https://arxiv.org/pdf/2210.10090.pdf"> pdf </a> 1011 | 1012 <a class="disabled" href="https://github.com/seva100/stylegan-for-facerec"> code </a> 1013 | 1014 <a href="https://www.youtube.com/watch?v=Bsi0RMTdEaI"> video </a> 1015 | 1016 <a href="https://seva100.github.io/stylegan-for-facerec"> page </a> 1017 </p> 1018 </div> 1019</div> 1020 1021<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 1022 <div class="col-sm-4 align-self-center"> 1023 <center> 1024 <img style="width:100%; padding: 10%; border-radius: 65%; "
1025 alt="depth estimation photo" src="/assets/images/people_counting.png"> 1026 </center> 1027 </div> 1028 <div class="col-sm-8 align-self-center"> 1029 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Leveraging Self-Supervision for Cross-Domain Crowd Counting</span></b></h6> 1030 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1031 Weizhe Liu, <b style="color: white">Nikita Durasov</b>, Pascal Fua<br> 1032 </p> 1033 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">CVPR 2022 (Oral)</span></p> 1034 <p style="text-align: center;"> 1035 1036 <a href="https://arxiv.org/abs/2103.16291"> pdf </a> 1037 | 1038 <a class="disabled" href="https://github.com/weizheliu/Cross-Domain-Crowd-Counting"> code </a> 1039 | 1040 <a href="https://www.youtube.com/watch?v=Lu30UmD5I9c"> video </a> 1041 | 1042 <a href="/projects/cross_domain_crowd_counting/"> page </a> 1043 </p> 1044 </div> 1045</div> 1046 1047 1048<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 1049 <div class="col-sm-4 align-self-center"> 1050 <center> 1051 <img style="width:100%; padding: 10%; border-radius: 65%; " 1052 alt="depth estimation photo" src="/assets/images/uncertainty_quantification.png"> 1053 </center> 1054 </div> 1055 <div class="col-sm-8 align-self-center"> 1056 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Masksembles for Uncertainty Estimation</span></b></h6> 1057 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1058 <b style="color: white">Nikita Durasov</b>, Timur Bagautdinov, Pierre Baque, Pascal Fua <br> 1059 </p> 1060 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">CVPR 2021</span></p> 1061 <p style="text-align: center;"> 1062 1063 <a href="https://arxiv.org/abs/2012.08334"> pdf </a> 1064 | 1065 <a href="https://github.com/nikitadurasov/masksembles"> code </a> 1066 | 1067 <a href="https://www.neuralconcept.com/post/the-importance-of-uncertainty-quantification-for-deep-learning-models-in-cae"> media </a> 1068 | 1069 <a href="https://www.youtube.com/watch?v=YWKVdn3kLp0"> video </a> 1070 | 1071 <a href="/projects/masksembles/"> page </a> 1072 </p> 1073 </div> 1074</div> 1075 1076<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 1077 <div class="col-sm-4 align-self-center"> 1078 <center> 1079 <img style="width:100%; padding: 10%; border-radius: 65%" 1080 alt="depth estimation photo" src="/assets/images/depth_sintel.png"> 1081 </center> 1082 </div> 1083 <div class="col-sm-8 align-self-center"> 1084 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Double refinement network for efficient indoor monocular depth estimation</span></b><h6> 1085 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1086 <b style="color: white">Nikita Durasov</b>, Mikhail Romanov, Valeriya Bubnova, Pavel Bogomolov, Anton Konushin <br> 1087 </p> 1088 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">IROS 2019</span></p> 1089 <p style="text-align: center; margin: 0px 20px 10px 20px;"> 1090 <a href="https://arxiv.org/abs/1811.08466"> pdf </a> 1091 | <a href="https://drive.google.com/file/d/1dOg7cJ6-Gf-ggaizbUU-l1w-njWwBVgT/view?usp=sharing"> video </a> 1092 </p> 1093 </div> 1094</div> 1095 1096<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 1097 <div class="col-sm-4 align-self-center"> 1098 <center> 1099 <img style="width:100%; padding: 10%; border-radius: 65%" 1100 alt="depth estimation photo" src="/assets/images/graph_square.png"> 1101 </center> 1102 </div> 1103 <div class="col-sm-8 align-self-center"> 1104 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">Learning node embeddings for influence set completion</span></b></h6> 1105 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1106 Sergei Ivanov, <b style="color: white">Nikita Durasov</b>, Evgeny Burnaev 1107 <br> 1108 </p> 1109 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">ICDM 2018</span></p> 1110 <p style="text-align: center;"> 1111 <a href="https://ieeexplore.ieee.org/document/8637545"> pdf </a> 1112 | 1113 <a href="https://github.com/nd7141/AWE"> code </a> 1114 </p> 1115 </div> 1116</div> 1117 1118<h2 class="divider line glow" contenteditable>Preprints</h2> 1119 1120<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 1121 <div class="col-sm-4 align-self-center"> 1122 <center> 1123 <img style="width:100%; padding: 10%; border-radius: 65%" 1124 alt="depth estimation photo" src="/assets/images/partal.png">
1125 </center> 1126 </div> 1127 <div class="col-sm-8 align-self-center"> 1128 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">PartAL: Efficient Partial Active Learning in Multi-Task Visual Settings</span></b></h6> 1129 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1130 <b style="color: white">Nikita Durasov</b>, Nik Dorndorf, Pascal Fua 1131 <br> 1132 </p> 1133 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">ARXIV 2022</span></p> 1134 <p style="text-align: center;"> 1135 <a href="https://arxiv.org/abs/2211.11546"> pdf </a> 1136 <!-- | 1137 - 1138 <a class="disabled" href="https://github.com/seva100/stylegan-for-facerec"> code </a> 1139 | 1140 - 1141 <a href="https://www.youtube.com/watch?v=Bsi0RMTdEaI"> video </a> 1142 | 1143 - 1144 <a href="https://seva100.github.io/stylegan-for-facerec"> page </a> --> 1145 </p> 1146 </div> 1147</div> 1148 1149<div class="row paper-item" style="margin: 20px 5px 20px 5px;"> 1150 <div class="col-sm-4 align-self-center"> 1151 <center> 1152 <img style="width:100%; padding: 10%; border-radius: 65%" 1153 alt="depth estimation photo" src="/assets/images/debosh1.png"> 1154 </center> 1155 </div> 1156 <div class="col-sm-8 align-self-center"> 1157 <h6 style="text-align: left; margin: 10px 20px 10px 20px;"><b><span class="title-highlights">DEBOSH: Deep Bayesian Shape Optimization</span></b></h6> 1158 <p style="font-size: 14px; margin: 0px 20px 10px 20px;"> 1159 <b style="color: white">Nikita Durasov</b>, Artem Lukoyanov, Jonathan Donier, Pascal Fua 1160 <br> 1161 </p> 1162 <p style="font-size: 15px; text-align: left; margin: 5px 20px 10px 20px;"><span class="topic-highlights">ARXIV 2021</span></p> 1163 <p style="text-align: center;"> 1164 <a href="https://arxiv.org/pdf/2109.13337.pdf"> pdf </a> 1165 | 1166 <a href="https://www.neuralconcept.com/post/debosh-deep-bayesian-shape-optimization"> media </a> 1167 | 1168 <a href="/projects/debosh_depricated/"> page </a> 1169 </p> 1170 </div> 1171</div> 1172 1173<h2 class="divider line glow" contenteditable>Invited Talks</h2> 1174 1175<hr style="border: 0.01px #363636 dashed; margin: 15px 10%;"> 1176 1177<ul> 1178 <li style="margin-bottom: 20px;"> 1179 <span class="topic-highlights">01/04/2025</span>: "Test-Time Training methods for Generalization" was presented at <span class="topic-highlights">AI Tinkerers</span> in Lausanne, Switzerland. <a href="https://torch-ttt.github.io/auto_examples/02_imagenet_tent.html">
1179 [demo] </a> 1180 </li> 1181 <li style="margin-bottom: 20px;"> 1182 <span class="topic-highlights">13/02/2025</span>: "Uncertainty and Generalization in CV" was presented at <span class="topic-highlights">LauzHack DL Bootcamp</span> in Lausanne, Switzerland. <a href="https://github.com/LauzHack/deep-learning-bootcamp/tree/winter25/day08"> [slides] </a> <a href="https://github.com/LauzHack/deep-learning-bootcamp/tree/winter25/day08"> [video] </a> 1183 </li> 1184 <li style="margin-bottom: 20px;"> 1185 <span class="topic-highlights">15/08/2024</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">Mila - Quebec AI Institute</span> in Montréal, Canada. 1186 </li> 1187 <li style="margin-bottom: 20px;"> 1188 <span class="topic-highlights">10/04/2024</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">UiT The Arctic University of Norway</span> in Tromsø, Norway. 1189 </li> 1190 <li style="margin-bottom: 20px;"> 1191 <span class="topic-highlights">12/02/2024</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">MBZUAI</span> in Abu Dhabi, UAE. 1192 </li> 1193 <li style="margin-bottom: 20px;"> 1194 <span class="topic-highlights">08/01/2024</span>: "Low-cost Uncertainty Estimation for Deep Learning" at <span class="topic-highlights">CVLAB Group Meeting</span> in Lausanne, Switzerland. 1195 </li> 1196 <li style="margin-bottom: 20px;"> 1197 <span class="topic-highlights">02/10/2023</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">Nordic AI Conference</span> in Copenhagen, Sweden. 1198 </li> 1199 <li style="margin-bottom: 20px;"> 1200 <span class="topic-highlights">17/08/2023</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">Stony Brook University</span> in New York, USA. 1201 </li> 1202 <li style="margin-bottom: 20px;"> 1203 <span class="topic-highlights">16/08/2023</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">NYU Courant Institute</span> in New York, USA. 1204 </li> 1205 <li style="margin-bottom: 20px;"> 1206 <span class="topic-highlights">09/08/2023</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">Apple</span>'s Park in Cupertino, USA. 1207 </li> 1208 <li style="margin-bottom: 20px;"> 1209 <span class="topic-highlights">08/08/2023</span>: "Deep Uncertainty" was presented at <span class="topic-highlights">Meta</span> in San Francisco, USA. 1210 </li> 1211 <li style="margin-bottom: 20px;"> 1212 <span class="topic-highlights">13/12/2022</span>: "Masksembles for Uncertainty Estimation" at <span class="topic-highlights">Uncertainty Reading Group</span> at <span class="topic-highlights">ETH</span> in Zurich, Switzerland. <a href="https://www.youtube.com/watch?v=dL5R7gzBPEc"> [video] </a> 1213 </li> 1214 <li style="margin-bottom: 20px;"> 1215 <span class="topic-highlights">30/09/2022</span>: "Enhanced Robustness of Transformer Detectors for Handling Omissions in Ground Truth Data" at <span class="topic-highlights">Apple Development Center</span> in Zurich, Switzerland. 1216 </li> 1217 <li style="margin-bottom: 20px;"> 1218 <span class="topic-highlights">26/08/2022</span>: "Single-shot Uncertainty Estimation" at <span class="topic-highlights">CVLAB Group Meeting</span> in Lausanne, Switzerland. 1219 </li> 1220 <li style="margin-bottom: 20px;"> 1221 <span class="topic-highlights">26/04/2022</span>: "Partial Active Learning with Incomplete Annotations" at <span class="topic-highlights">VILAB Group Meeting</span> in Lausanne, Switzerland. 1222 </li> 1223 <li style="margin-bottom: 20px;"> 1224 <span class="topic-highlights">20/12/2021</span>: "Robust Prediction of Stereo Disparity Under Uncertainty" at <span class="topic-highlights">Amazon Prime Air</span> in Graz, Austria. 1225 </li> 1226 <li style="margin-bottom: 20px;"> 1227 <span class="topic-highlights">22/06/2021</span>: "Masksembles for Uncertainty Estimation" at <span class="topic-highlights">CVPR 2021</span>. <a href="https://www.youtube.com/watch?v=YWKVdn3kLp0"> [video] </a> <a href="https://arxiv.org/abs/2012.08334"> [paper] </a> 1228 </li> 1229</ul> 1230 1231<hr style="border: 0.01px #363636 dashed; margin: 15px 10%;"> 1232 1233<!--
1233<script> 1234 var ctx = document.getElementById('citationChart').getContext('2d'); 1235 1236 // Create gradient 1237 var gradient = ctx.createLinearGradient(0, 0, 0, 300); 1238 gradient.addColorStop(0, 'rgba(61, 14, 85, 1)'); // Light blue 1239 gradient.addColorStop(1, 'rgba(17, 66, 154, 1)'); // Dark blue 1240 1241 var citationChart = new Chart(ctx, { 1242 type: 'bar', 1243 data: { 1244 labels: ['2019', '2020', '2021', '2022', '2023', '2024', '2025'], // Add your years here 1245 datasets: [{ 1246 label: 'Citations', 1247 data: [5, 2, 20, 33, 77, 90, 60], // Add your citation numbers here 1248 backgroundColor: gradient, // Use the gradient for all bars 1249 borderColor: [ 1250 'rgba(255, 255, 255, 0.2)', 1251 'rgba(255, 255, 255, 0.2)', 1252 'rgba(255, 255, 255, 0.2)', 1253 'rgba(255, 255, 255, 0.2)', 1254 'rgba(255, 255, 255, 0.2)' 1255 ], 1256 borderWidth: 1 1257 }] 1258 }, 1259 options: { 1260 scales: { 1261 y: { 1262 beginAtZero: true, 1263 grid: { 1264 color: 'rgba(255, 255, 255, 0.2)', 1265 borderWidth: 1, 1266 drawBorder: false, 1267 } 1268 }, 1269 x: { 1270 grid: { 1271 color: 'rgba(255, 255, 255, 0.2)', 1272 borderWidth: 1, 1273 drawBorder: false, 1274 } 1275 } 1276 }, 1277 plugins: { 1278 legend: { 1279 display: true 1280 } 1281 } 1282 } 1283 }); 1284 </script>
1284 --> 1285
1286<script> 1287 const ctx = document.getElementById('citationChart').getContext('2d'); 1288 1289 // Create gradient 1290 const gradient = ctx.createLinearGradient(0, 0, 0, 300); 1291 gradient.addColorStop(0, 'rgba(61, 14, 85, 1)'); 1292 gradient.addColorStop(1, 'rgba(17, 66, 154, 1)'); 1293 1294 fetch('/assets/citations.json') 1295 .then(response => response.json()) 1296 .then(data => { 1297 const citationsPerYear = data.citations_per_year; 1298 const totalCitations = data.total_citations; 1299 const hIndex = data.h_index; 1300 const i10Index = data.i10_index; 1301 1302 // 1. Update numbers in stats list 1303 document.querySelector('.my-stats-list').innerHTML = ` 1304 <li class="list-inline-item"><span class="topic-highlights">Citations: ${totalCitations}</span></li> 1305 <li class="list-inline-item"><span class="topic-highlights">h-index: ${hIndex}</span></li> 1306 <li class="list-inline-item"><span class="topic-highlights">i10-index: ${i10Index}</span></li> 1307 `; 1308 1309 const years = Object.keys(citationsPerYear).sort(); 1310 const citations = years.map(year => citationsPerYear[year]); 1311 1312 // Compute projection for the last (current) year 1313 const now = new Date(); 1314 const startOfYear = new Date(now.getFullYear(), 0, 1); 1315 const daysPassed = Math.floor((now - startOfYear) / (1000 * 60 * 60 * 24)) + 1; 1316 const lastYearCitations = citations[citations.length - 1]; 1317 const projectedValue = Math.round(lastYearCitations / daysPassed * 365); 1318 1319 const projectionData = [...Array(years.length - 1).fill(null), projectedValue]; 1320 1321 const previousTotal = citations.slice(0, -1).reduce((a, b) => a + b, 0); 1322 const projectedTotal = previousTotal + projectedValue; 1323 1324 // Gradient for projection bar 1325 const projGradient = ctx.createLinearGradient(0, 0, 0, 300); 1326 projGradient.addColorStop(0, 'rgba(200, 100, 50, 0.3)'); 1327 projGradient.addColorStop(1, 'rgba(230, 160, 60, 0.3)'); 1328 1329 const citationChart = new Chart(ctx, { 1330 type: 'bar', 1331 data: { 1332 labels: years, 1333 datasets: [ 1334 { 1335 label: `Projection (total=${projectedTotal})`, 1336 data: projectionData, 1337 backgroundColor: projGradient, 1338 borderColor: Array(years.length).fill('rgba(255, 255, 255, 0.2)'), 1339 borderWidth: 1, 1340 grouped: false, 1341 order: 1 1342 }, 1343 { 1344 label: 'Citations', 1345 data: citations, 1346 backgroundColor: gradient, 1347 borderColor: Array(years.length).fill('rgba(255, 255, 255, 0.2)'), 1348 borderWidth: 1, 1349 grouped: false, 1350 order: 0 1351 } 1352 ] 1353 }, 1354 options: { 1355 scales: { 1356 y: { 1357 beginAtZero: true, 1358 grid: { 1359 color: 'rgba(255, 255, 255, 0.2)', 1360 borderWidth: 1, 1361 drawBorder: false, 1362 } 1363 }, 1364 x: { 1365 grid: { 1366 color: 'rgba(255, 255, 255, 0.2)', 1367 borderWidth: 1, 1368 drawBorder: false, 1369 } 1370 } 1371 }, 1372 plugins: { 1373 legend: { 1374 display: true 1375 } 1376 } 1377 } 1378 }); 1379 }); 1380</script>
1380 1381
1382<script> 1383fetch("/assets//views.json") 1384 .then((res) => res.json()) 1385 .then((data) => { 1386 const totalViews = data["/*"] || 0; 1387 document.getElementById("ga_site_pv").textContent = totalViews.toLocaleString(); 1388 }) 1389 .catch((err) => { 1390 console.error("Failed to load total view count", err); 1391 document.getElementById("ga_site_pv").textContent = "N/A"; 1392 }); 1393</script>
1393 1394 </div> 1395 </div> 1396 </main><footer class="site-footer h-card"> 1397 <data class="u-url" href="/"></data> 1398 1399 <div class="wrapper"> 1400 1401 <div class="footer-col-wrapper"> 1402 <div class="footer-col"> 1403 <!-- <p class="feed-subscribe"> 1404 <a href="/feed.xml"> 1405 <svg class="svg-icon orange"> 1406 <use xlink:href="/assets/minima-social-icons.svg#rss"></use> 1407 </svg><span>Subscribe</span> 1408 </a> 1409 </p> --> 1410 </div> 1411 <div class="footer-col"> 1412 <p style="opacity: 0.2">Thoughts on deep learning, tech, and whatever else seems interesting.</p> 1413 </div> 1414 </div> 1415 1416 <div class="social-links"><style> 1417 .google-scholar-icon { 1418 transform: rotate(180deg); 1419 transform-origin: center; 1420 display: inline-block; 1421 } 1422</style> 1423 1424<ul class="social-media-list"> 1425 <!--<li><a rel="me" href="https://github.com/nikitadurasov" class="svg-a" title="nikitadurasov"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#github"></use></svg></a></li><li><a rel="me" href="https://www.linkedin.com/in/durasovnikita" class="svg-a" title="durasovnikita"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#linkedin"></use></svg></a></li><li><a rel="me" href="https://twitter.com/norange_ai" class="svg-a" title="norange_ai"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#twitter"></use></svg></a></li>--><li><a rel="me" href="https://github.com/nikitadurasov" class="svg-a" title="nikitadurasov"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#github"></use></svg></a></li><li><a rel="me" href="https://www.linkedin.com/in/durasovnikita" class="svg-a" title="durasovnikita"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#linkedin"></use></svg></a></li><li><a rel="me" href="https://twitter.com/norange_ai" class="svg-a" title="norange_ai"><svg class="svg-icon blue"><use xlink:href="/assets/minima-social-icons.svg#twitter"></use></svg></a></li><!----><li> 1426 <a rel="me" href="https://scholar.google.com/citations?user=KMMvTfcAAAAJ%26hl" class="svg-a" title="Google Scholar Profile for KMMvTfcAAAAJ&hl"> 1427 <svg class="svg-icon blue google-scholar-icon"> 1428 <use xlink:href="/assets/minima-social-icons.svg#google-scholar"></use> 1429 </svg> 1430 </a> 1431 </li></ul> 1432 1433 1434</div> 1435 1436 </div> 1437 1438</footer> 1439<!-- jQuery first, then Popper.js, then Bootstrap JS. --> 1440
1440<script src="https://code.jquery.com/jquery-3.3.1.slim.min.js" integrity="sha384-q8i/X+965DzO0rT7abK41JStQIAqVgRVzpbzo5smXKp4YfRvH+8abtTE1Pi6jizo" crossorigin="anonymous"></script>
1440 1441
1441<script src="https://cdnjs.cloudflare.com/ajax/libs/popper.js/1.14.7/umd/popper.min.js" integrity="sha384-UO2eT0CpHqdSJQ6hJty5KVphtPhzWj9WO1clHTMGa3JDZwrnQq4sF86dIHNDz0W1" crossorigin="anonymous"></script>
1441 1442
1442<script src="https://stackpath.bootstrapcdn.com/bootstrap/4.3.1/js/bootstrap.min.js" integrity="sha384-JjSmVgyd0p3pXB1rRibZUAYoIIy6OrQ6VrjIEaFf/nJGzIxFDsf4x0xIM+B07jRM" crossorigin="anonymous"></script>
1442 1443 1444 </body> 1445 1446</html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.