-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathresearch.html
More file actions
311 lines (297 loc) · 20 KB
/
Copy pathresearch.html
File metadata and controls
311 lines (297 loc) · 20 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width,initial-scale=1" />
<link rel="icon" type="image/svg+xml" href="assets/img/favicon.svg" />
<title>Research — VIDAL Lab</title>
<link rel="preconnect" href="https://fonts.googleapis.com" />
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
<link href="https://fonts.googleapis.com/css2?family=Instrument+Serif:ital@0;1&family=Space+Grotesk:wght@300;400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet" />
<link rel="stylesheet" href="assets/css/styles.css" />
<link rel="stylesheet" href="assets/css/subpages.css" />
<link rel="stylesheet" href="assets/css/research-topics.css" />
<script defer src="assets/js/nav.js"></script>
</head>
<body data-page="research">
<header class="nav" id="nav">
<a class="brand" href="index.html">
<span class="brand-mark" aria-hidden="true">
<svg viewBox="0 0 40 40" width="28" height="28" aria-hidden="true">
<g fill="#011F5B" opacity="0.18">
<rect x="9" y="1" width="7" height="7"/><rect x="17" y="1" width="7" height="7"/><rect x="25" y="1" width="7" height="7"/><rect x="33" y="1" width="7" height="7"/>
<rect x="1" y="9" width="7" height="7"/><rect x="17" y="9" width="7" height="7"/><rect x="25" y="9" width="7" height="7"/><rect x="33" y="9" width="7" height="7"/>
<rect x="1" y="17" width="7" height="7"/><rect x="9" y="17" width="7" height="7"/><rect x="25" y="17" width="7" height="7"/><rect x="33" y="17" width="7" height="7"/>
<rect x="1" y="25" width="7" height="7"/><rect x="9" y="25" width="7" height="7"/><rect x="17" y="25" width="7" height="7"/><rect x="33" y="25" width="7" height="7"/>
<rect x="1" y="33" width="7" height="7"/><rect x="9" y="33" width="7" height="7"/><rect x="17" y="33" width="7" height="7"/><rect x="25" y="33" width="7" height="7"/>
</g>
<g fill="#990000">
<rect x="1" y="1" width="7" height="7"/><rect x="9" y="9" width="7" height="7"/><rect x="17" y="17" width="7" height="7"/><rect x="25" y="25" width="7" height="7"/><rect x="33" y="33" width="7" height="7"/>
</g>
</svg>
</span>
<span class="brand-word"><em>Vidal</em>Lab<sup>↗</sup></span>
</a>
<nav class="nav-links" aria-label="Primary">
<a href="index.html" data-key="home"><span>01</span> Home</a>
<a href="people.html" data-key="people"><span>02</span> People</a>
<a href="research.html" data-key="research" class="active"><span>03</span> Research</a>
<div class="nav-more">
<button class="nav-more-btn" aria-haspopup="true" aria-expanded="false">
<span>04</span> More
<svg viewBox="0 0 24 24" width="12" height="12" fill="none" stroke="currentColor" stroke-width="1.8"><path d="M6 9l6 6 6-6"/></svg>
</button>
<div class="nav-more-menu" role="menu">
<a href="teaching.html" data-key="teaching" role="menuitem"><span>a</span> Teaching</a>
<a href="talks.html" data-key="talks" role="menuitem"><span>b</span> Talks</a>
<a href="tutorials.html" data-key="tutorials" role="menuitem"><span>c</span> Tutorials</a>
<a href="data-code.html" data-key="data" role="menuitem"><span>d</span> Data & Code</a>
</div>
</div>
</nav>
</header>
<section class="research-topics-section" style="margin-top:140px">
<header class="section-head">
<div class="section-num">§ 03·1</div>
<h2>Research <em>areas</em> & topics</h2>
<p class="section-note">Click an area on the left; click a topic on the right to filter the publication list below.</p>
</header>
<div class="areas-topics">
<div class="areas-list">
<article class="area area-tab active" data-category="ml">
<div class="area-num">i.</div>
<div class="area-body">
<h3>Machine <em>Learning</em></h3>
<p>Deep learning theory, trustworthy AI, generative models, parsimonious representation learning, continual learning, and optimization for non-convex loss landscapes.</p>
<ul class="area-tags"><li>Deep theory</li><li>Trustworthy</li><li>Generative</li><li>Optimization</li><li>Continual</li></ul>
</div>
</article>
<article class="area area-tab" data-category="cv">
<div class="area-num">ii.</div>
<div class="area-body">
<h3>Computer <em>Vision</em></h3>
<p>Vision–language models, human motion understanding, video analysis, image analysis, and geometric 3D vision — with an emphasis on interpretability and generalization.</p>
<ul class="area-tags"><li>VLMs</li><li>Motion</li><li>3D</li><li>Video</li><li>Interpretable</li></ul>
</div>
</article>
<article class="area area-tab" data-category="dyn">
<div class="area-num">iii.</div>
<div class="area-body">
<h3>Dynamics & <em>Robotics</em></h3>
<p>Methods for analyzing and learning linear, hybrid, and multi-agent systems — from theoretical guarantees to deployment on physical platforms.</p>
<ul class="area-tags"><li>Hybrid</li><li>Multi-agent</li><li>Control</li><li>Safety</li></ul>
</div>
</article>
<article class="area area-tab" data-category="med">
<div class="area-num">iv.</div>
<div class="area-body">
<h3>AI for <em>Healthcare</em></h3>
<p>Translating advances in AI into clinical systems for radiology, cardiology, neurology, hematology, and surgery — together with clinicians at Penn Medicine and beyond.</p>
<ul class="area-tags"><li>Radiology</li><li>Cardiology</li><li>Surgery</li><li>Neurology</li></ul>
</div>
</article>
</div>
<div class="topics-list">
<div class="research-category-panel active" id="panel-ml">
<div class="research-topic-item">
<h3>Deep Learning Theory</h3>
<button class="view-pubs-btn" data-topics="Deep Learning">View Publications →</button>
<p>We study the mathematical principles behind modern neural networks. Our lab focuses on understanding the optimization landscape and generalization properties of positively homogeneous networks, analyzing the learning dynamics and implicit bias of gradient-based methods, and providing principled explanations to deep learning phenomena such as neural collapse, low-rank adaptation, and learning on the edge of stability.</p>
</div>
<div class="research-topic-item">
<h3>Trustworthy AI</h3>
<button class="view-pubs-btn" data-topics="Adversarial Robustness,Interpretable AI">View Publications →</button>
<p>We develop AI systems that are interpretable, robust, and reliable for high-stakes applications. Our information pursuit framework enables explainable-by-design models that make predictions from a sequence of informative question–answer pairs. We also analyze when robust classifiers are guaranteed to exist and design adversarial attacks that stress-test visual classifiers and large language models.</p>
</div>
<div class="research-topic-item">
<h3>Deep Generative Models</h3>
<button class="view-pubs-btn" data-topics="Deep Learning,Computer Vision">View Publications →</button>
<p>We explore the theory and practice of deep generative models — convergence guarantees for generative-model inversion, transformer-based diffusion, flow matching aligned with the multimodal structure of real data, and accurate diffusion-based image restoration.</p>
</div>
<div class="research-topic-item">
<h3>Parsimonious Representation Learning</h3>
<button class="view-pubs-btn" data-topics="GPCA,Sparse Optimization,Manifold clustering,Boosting">View Publications →</button>
<p>We focus on discovering sparse and low-rank structure in high-dimensional data — Generalized PCA (GPCA), Sparse Subspace Clustering (SSC), Dual Principal Component Pursuit (DPCP), and nonconvex matrix and tensor factorization with global-optimality guarantees.</p>
</div>
<div class="research-topic-item">
<h3>Continual Learning</h3>
<button class="view-pubs-btn" data-topics="Continual Learning">View Publications →</button>
<p>We build models that learn tasks sequentially without forgetting. Frameworks like the Ideal Continual Learner (ICL) and LoRanPAC unify existing methods, provide theoretical guarantees, and maintain stability over long task sequences.</p>
</div>
<div class="research-topic-item">
<h3>Optimization</h3>
<button class="view-pubs-btn" data-topics="Optimization">View Publications →</button>
<p>We analyze convergence and global optimality of optimization algorithms for machine learning — subspace clustering, nonconvex matrix and tensor factorization, dictionary learning, generative-model inversion, and deep-learning training. We also study distributed optimization on Riemannian manifolds and accelerated methods.</p>
</div>
</div>
<div class="research-category-panel" id="panel-cv">
<div class="research-topic-item">
<h3>Vision–Language Models</h3>
<button class="view-pubs-btn" data-topics="Computer Vision,Interpretable AI">View Publications →</button>
<p>We develop vision–language models that connect visual data with semantic concepts, including concept-based interpretable classification, knowledge-pursuit prompting for zero-shot multimodal synthesis, and concept-based image-editing methods.</p>
</div>
<div class="research-topic-item">
<h3>Human Motion Analysis</h3>
<button class="view-pubs-btn" data-topics="Computer Vision">View Publications →</button>
<p>We develop methods for understanding human motion from video and skeletal data — action recognition, detection, segmentation, pose estimation, and multimodal motion representations supporting recognition, retrieval, and generation.</p>
</div>
<div class="research-topic-item">
<h3>Video Analysis</h3>
<button class="view-pubs-btn" data-topics="Dynamic textures,2D motion segmentation,3D motion segmentation">View Publications →</button>
<p>We model dynamic visual phenomena in video — dynamic textures via dynamical systems, semantic video segmentation, robust video classification, and perpetual generation of dynamic scenes with 3D consistency.</p>
</div>
<div class="research-topic-item">
<h3>Image Analysis</h3>
<button class="view-pubs-btn" data-topics="Computer Vision,nonrigid shape,GPCA">View Publications →</button>
<p>Algorithms for image segmentation, alpha matting, semantic segmentation with structured outputs, visual representation learning, and image deblurring with both classical and diffusion-based priors.</p>
</div>
<div class="research-topic-item">
<h3>Geometric Vision</h3>
<button class="view-pubs-btn" data-topics="3D motion segmentation,2D motion segmentation,omnidirectional vision">View Publications →</button>
<p>Geometric methods for recovering 3D structure and motion — multibody motion segmentation, the Hopkins 155 benchmark, robust object pose estimation, and optimization-based geometric estimation via DPCP and IRLS.</p>
</div>
</div>
<div class="research-category-panel" id="panel-dyn">
<div class="research-topic-item">
<h3>Learning and Analyzing Linear, Bilinear and Hybrid Dynamical Systems</h3>
<button class="view-pubs-btn" data-topics="Hybrid systems,Hybrid system identification,Realization theory">View Publications →</button>
<p>Conditions for observability and algorithms for system identification of hybrid systems, realization theory for stochastic jump-Markov and bilinear systems, and identification of linear systems with sparse inputs.</p>
</div>
<div class="research-topic-item">
<h3>Geometry and Distances of Spaces of Dynamical Systems</h3>
<button class="view-pubs-btn" data-topics="Distances,Dynamic textures">View Publications →</button>
<p>Principled distances and similarity measures for comparing dynamical systems — Binet–Cauchy kernels, group-action-induced distances, and alignment distances, with strong impact on video comparison and action recognition.</p>
</div>
<div class="research-topic-item">
<h3>Dynamical-Systems Perspectives on Optimization</h3>
<button class="view-pubs-btn" data-topics="Optimization">View Publications →</button>
<p>Optimization through the lens of dynamical systems — accelerated gradient descent, ADMM, conformal symplectic and relativistic optimization, nonsmooth dynamical systems, and proximal splitting methods.</p>
</div>
<div class="research-topic-item">
<h3>Distributed Optimization and Consensus on Manifolds</h3>
<button class="view-pubs-btn" data-topics="formation control,Distances,Optimization">View Publications →</button>
<p>Geometric methods for distributed optimization and consensus — Riemannian motion estimation on the essential manifold, Riemannian centers of mass, and consensus algorithms with theoretical guarantees.</p>
</div>
<div class="research-topic-item">
<h3>Robotics and Autonomous Systems</h3>
<button class="view-pubs-btn" data-topics="formation control,Computer Vision">View Publications →</button>
<p>Geometric and control-theoretic methods for autonomous robots — vision-based control for autonomous helicopter landing and game-theoretic pursuit–evasion strategies.</p>
</div>
</div>
<div class="research-category-panel" id="panel-med">
<div class="research-topic-item">
<h3>Diffusion MRI Analysis</h3>
<button class="view-pubs-btn" data-topics="Biomedical">View Publications →</button>
<p>HARDI reconstruction, restoration, registration, and segmentation via information geometry, sparse representation, and non-convex optimization — enabling biomarker discovery for neurological disorders.</p>
</div>
<div class="research-topic-item">
<h3>Surgical Activity Recognition and Skill Assessment</h3>
<button class="view-pubs-btn" data-topics="Biomedical,Computer Vision">View Publications →</button>
<p>Bag-of-spatiotemporal features, sparse HMMs, CRFs, and spatio-temporal deep architectures for surgical-workflow understanding — plus the JHU-ISI gesture and skill assessment benchmark.</p>
</div>
<div class="research-topic-item">
<h3>Stem Cell-Derived Cardiomyocytes</h3>
<button class="view-pubs-btn" data-topics="Biomedical">View Publications →</button>
<p>Clustering and classification of cardiac cells based on cell morphology and contractile dynamics — shape analysis via metamorphosis models and recurrent networks for temporal contraction patterns.</p>
</div>
<div class="research-topic-item">
<h3>Computational Microscopy & Point-of-Care Diagnostics</h3>
<button class="view-pubs-btn" data-topics="Biomedical,Sparse Optimization">View Publications →</button>
<p>Holographic reconstruction, hybrid physics+deep phase retrieval, and encoder–decoder cell-detection networks for point-of-care diagnostics — automated CBC and lensless urinalysis.</p>
</div>
<div class="research-topic-item">
<h3>Vision for Neurological & Developmental Disorders</h3>
<button class="view-pubs-btn" data-topics="Biomedical,Computer Vision,Interpretable AI">View Publications →</button>
<p>Computer-vision methods for analyzing human movement — infant action recognition, motor-tic detection for Tourette assessment, and the CAMI motor-imitation framework as a biomarker for autism.</p>
</div>
<div class="research-topic-item">
<h3>Medical Vision–Language & Foundation Models</h3>
<button class="view-pubs-btn" data-topics="Biomedical,Interpretable AI">View Publications →</button>
<p>Multimodal AI combining medical images and clinical text — extracting structured medical facts from radiology reports, grounding clinical concepts in chest X-rays, and building interpretable foundation models for radiology and cardiology.</p>
</div>
</div>
</div>
</div>
<div style="display:flex; justify-content:center; margin-top:48px;">
<a class="btn btn-ghost btn-lg" href="https://scholar.google.com/citations?user=uT8fgagAAAAJ" target="_blank" rel="noopener">
Full publication list on Google Scholar
<svg viewBox="0 0 24 24" width="18" height="18" fill="none" stroke="currentColor" stroke-width="1.6"><path d="M7 17L17 7M9 7h8v8"/></svg>
</a>
</div>
</section>
<section id="publications-section" class="rt-section" style="margin-top:80px">
<header class="section-head">
<div class="section-num">§ 03·2</div>
<h2>Publications by <em>topic</em></h2>
<p class="section-note">Select one or more topics to filter the list.</p>
</header>
<p class="topic-hint">Click a topic chip to filter; click again to remove. "All" clears.</p>
<div id="topic-tabs" class="topic-tabs"></div>
<p id="pub-count" class="pub-count"></p>
<div id="publications-container"></div>
</section>
<section style="margin: 100px 0 0;" class="affiliations">
<div style="margin: 0 var(--pad-x);">
<span class="aff-label">Affiliated centers</span>
<div class="aff-marquee" style="margin-top: 24px;">
<div class="aff-track">
<span>IDEAS Center</span><span class="aff-dot"></span>
<span>ASSET Center</span><span class="aff-dot"></span>
<span>GRASP Lab</span><span class="aff-dot"></span>
<span>Penn Engineering</span><span class="aff-dot"></span>
<span>Department of ESE</span><span class="aff-dot"></span>
<span>Penn Medicine</span><span class="aff-dot"></span>
<span>IDEAS Center</span><span class="aff-dot"></span>
<span>ASSET Center</span><span class="aff-dot"></span>
<span>GRASP Lab</span><span class="aff-dot"></span>
<span>Penn Engineering</span><span class="aff-dot"></span>
</div>
</div>
</div>
</section>
<section class="cta" style="margin-top:120px;">
<div class="cta-inner">
<h2 class="cta-title">Looking to <em>collaborate</em>?</h2>
<p class="cta-note">We welcome collaborations with clinicians, scientists, and engineers tackling problems where mathematical foundations matter.</p>
<div class="cta-actions">
<a class="btn btn-primary btn-lg" href="mailto:vidalr@seas.upenn.edu">Email Prof. Vidal <svg viewBox="0 0 24 24" width="18" height="18" fill="none" stroke="currentColor" stroke-width="1.6"><path d="M5 12h14M13 6l6 6-6 6"/></svg></a>
<a class="btn btn-ghost btn-lg" href="people.html">See the team</a>
</div>
</div>
</section>
<footer class="foot">
<div class="foot-mega" aria-hidden="true">VIDAL</div>
<div class="foot-grid">
<div class="foot-col">
<div class="foot-h">Vidal Lab</div>
<p>Vision, Dynamics & Learning<br/>University of Pennsylvania<br/>Department of Electrical & Systems Engineering</p>
<p class="mono">Philadelphia, PA</p>
</div>
<div class="foot-col">
<div class="foot-h">Sitemap</div>
<ul>
<li><a href="index.html">Home</a></li>
<li><a href="people.html">People</a></li>
<li><a href="research.html">Research</a></li>
<li><a href="teaching.html">Teaching</a></li>
<li><a href="talks.html">Talks</a></li>
<li><a href="tutorials.html">Tutorials</a></li>
<li><a href="data-code.html">Data & Code</a></li>
</ul>
</div>
<div class="foot-col">
<div class="foot-h">Contact</div>
<p>Prof. René Vidal<br/><a href="mailto:vidalr@seas.upenn.edu">vidalr@seas.upenn.edu</a></p>
</div>
<div class="foot-col">
<div class="foot-h">Join the Lab</div>
<p>We seek students at the boundary of mathematical foundations and impactful AI. PhD applications go through Penn ESE; postdoc and visitor inquiries direct to Prof. Vidal.</p>
<p><a href="mailto:vidalr@seas.upenn.edu">Email Prof. Vidal →</a></p>
<p class="mono dim">© 2026 Vidal Lab</p>
</div>
</div>
</footer>
<script src="assets/js/research-topics.js"></script>
<script src="assets/js/publications.js"></script>
</body>
</html>