-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.html
More file actions
820 lines (664 loc) · 42.8 KB
/
Copy pathindex.html
File metadata and controls
820 lines (664 loc) · 42.8 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
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Artificial Intelligence for Medicine and Science - Zitnik Lab</title>
<link rel="stylesheet" href="/assets/css/app.css">
<link rel="shortcut icon" type="image/png"
href="/favicon.png"
/>
<script defer src="https://use.fontawesome.com/releases/v5.3.1/js/all.js"></script>
<!-- Begin Jekyll SEO tag v2.8.0 -->
<title>Artificial Intelligence for Medicine and Science | Zitnik Lab</title>
<meta name="generator" content="Jekyll v4.3.4" />
<meta property="og:title" content="Artificial Intelligence for Medicine and Science" />
<meta name="author" content="Marinka Zitnik" />
<meta property="og:locale" content="en_US" />
<meta name="description" content="Artificial Intelligence (AI), Machine Learning (ML), Medicine, Science, and Drug Discovery at Harvard" />
<meta property="og:description" content="Artificial Intelligence (AI), Machine Learning (ML), Medicine, Science, and Drug Discovery at Harvard" />
<link rel="canonical" href="https://zitniklab.hms.harvard.edu/" />
<meta property="og:url" content="https://zitniklab.hms.harvard.edu/" />
<meta property="og:site_name" content="Zitnik Lab" />
<meta property="og:type" content="website" />
<meta name="twitter:card" content="summary" />
<meta property="twitter:title" content="Artificial Intelligence for Medicine and Science" />
<meta name="twitter:site" content="@marinkazitnik" />
<meta name="twitter:creator" content="@Marinka Zitnik" />
<script type="application/ld+json">
{"@context":"https://schema.org","@type":"WebSite","author":{"@type":"Person","name":"Marinka Zitnik"},"description":"Artificial Intelligence (AI), Machine Learning (ML), Medicine, Science, and Drug Discovery at Harvard","headline":"Artificial Intelligence for Medicine and Science","name":"Zitnik Lab","url":"https://zitniklab.hms.harvard.edu/"}</script>
<!-- End Jekyll SEO tag -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-162129505-1"></script>
<script>
window['ga-disable-UA-162129505-1'] = window.doNotTrack === "1" || navigator.doNotTrack === "1" || navigator.doNotTrack === "yes" || navigator.msDoNotTrack === "1";
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-162129505-1');
</script><!-- head scripts --></head>
<body>
<nav class="navbar is-primary" >
<div class="container">
<div class="navbar-brand">
<a href="/" class="navbar-item"><b>
Zitnik Lab
</b></a>
<a role="button" class="navbar-burger burger" aria-label="menu" aria-expanded="false" data-target="navMenu">
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
</a>
</div>
<div class="navbar-menu" id="navMenu">
<div class="navbar-start">
<!-- <a href="/" class="navbar-item is-active"><b>Home</b></a>-->
<div class="navbar-item has-dropdown is-hoverable">
<a href="/#" class="navbar-link "><b>About</b></a>
<div class="navbar-dropdown">
<a href="/bio/" class="navbar-item "><b>Bio</b></a>
<a href="/contact/" class="navbar-item "><b>Contact</b></a>
<a href="/talks/" class="navbar-item "><b>Recent Talks</b></a>
</div>
</div>
<a href="/research/" class="navbar-item "><b>Research</b></a>
<a href="/publications/" class="navbar-item "><b>Publications</b></a>
<a href="/people/" class="navbar-item "><b>Members</b></a>
<a href="/meetings/" class="navbar-item "><b>Teaching & Conferences</b></a>
<a href="/data/" class="navbar-item "><b>Datasets</b></a>
<a href="/software/" class="navbar-item "><b>AI Models</b></a>
<a href="https://aiscientist.tools/" class="navbar-item "><b>ToolUniverse</b></a>
<a href="https://tdcommons.ai/" class="navbar-item "><b>TDC</b></a>
<a href="https://www.openscientist.ai/" class="navbar-item "><b>Open AI Scientists</b></a>
<a href="/news/" class="navbar-item "><b>News</b></a>
<a href="/jobs/" class="navbar-item "><b>Join Us</b></a>
</div>
</div>
</div>
</nav>
<section class="hero is-medium is-bold is-primary" style="background: url('/hero.jpg') no-repeat center center; background-size: cover;" >
<div class="hero-body">
<div class="container">
<p class="title is-2">Artificial Intelligence for Medicine and Science</p>
<p class="subtitle is-3"><br/></p>
<a href="/jobs" class="button is-info is-large">Open Positions</a>
</div>
</div>
</section>
<section class="section">
<div class="container">
<div class="columns">
<div class="column is-8">
<div class="content">
<div class="box has-background-primary has-text-white">
<h4 class="has-text-white">Machine Learning Foundations and Biomedical Discovery</h4>
<p>
<b>
Our overarching goal is to establish the foundations of AI for science and medicine through systems that generate new biological insight and are capable of accelerating the discovery, design, and development of treatments for the most pressing medical challenges of our time.
</b>
</p>
<p>
We develop artificial intelligence and machine learning models, with emphasis on systems that are grounded in knowledge. Our work includes multimodal, generative and agentic AI models trained that can reason and empower scientific discovery across molecular, cellular, tissue, and patient scales.
</p>
</div>
<div class="box has-background-link has-text-white">
<h4 class="has-text-white"><a href="https://www.openscientist.ai/" class="has-text-white has-text-weight-bold">Open AI Scientists Initiative | Join Us</a></h4>
<p>
AI scientists are autonomous AI systems that reason, hypothesize, and experiment alongside human researchers. <a href="https://www.openscientist.ai/" class="has-text-white has-text-weight-bold">Through our Open AI Scientists Initiative,</a> we develop and study AI scientists that integrate research and clinical data with scientific tools to generate new hypotheses, prioritize experiments, and produce scientific insight. We evaluate these systems through discovery loops that connect AI reasoning with experiments in biological and clinical labs.
</p>
</div>
<div class="box has-background-info has-text-white">
<h4 class="has-text-white">AI for Science | Scientific Discovery and Therapeutic Science</h4>
<p>
Science is entering a new era in which AI can do more than analyze data: it can help generate hypotheses, connect disparate evidence, guide experiments, and accelerate discovery. Biological systems are deeply interconnected across scales, from genes and proteins to cells, tissues, and whole organisms. Our research develops AI methods that learn across these levels of organization to uncover biological principles, model disease mechanisms, and advance the design of therapeutics and other interventions.
<!-- For centuries, the method of discovery—the fundamental practice of science that scientists use to explain the natural world systematically and logically—has remained largely the same. We are using AI to change that. The natural world is interconnected, from the various facets of genome regulation to the molecular and organismal levels. These interactions across different levels yield a bewildering degree of complexity. Our research seeks to disentangle this complexity, developing AI models that advance drug design and help develop new kinds of therapies. -->
</p>
</div>
<div class="box has-background-info has-text-white">
<h4 class="has-text-white">AI for Medicine | Individualized Diagnosis and Treatment</h4>
<p>
Human health can now be measured with unprecedented precision through genomic data, cellular atlases, molecular profiles, clinical records, imaging, and treatment histories. The central challenge is to integrate and reason over these heterogeneous data to support diagnosis, prognosis, and treatment selection. Our research builds AI systems that combine patient data with biomedical knowledge to enable more precise, evidence-based, and individualized medicine, with the goal of matching each patient to the right treatment and improving consistency between clinical outcomes and biological insight.
<!-- The state of a person is described with increasing precision incorporating modalities like genetic code, cellular atlases, molecular datasets, and therapeutics—the challenge is how to reason over these data to develop powerful disease diagnostics and empower new kinds of therapies. Our research creates new avenues for fusing knowledge and patient data to give the right patient the right treatment at the right time and have medicinal effects that are consistent from person to person and with results in the laboratory. -->
</p>
</div>
<style>
/* CSS for the video container */
.video-container {
position: relative;
width: 100%;
/* 16:9 aspect ratio padding-bottom */
padding-bottom: 56.25%;
height: 0;
overflow: hidden;
border-radius: 20px;
box-shadow: 0 4px 12px rgba(0, 0, 0, 0.15);
}
/* CSS for the iframe inside the container */
.video-container iframe {
position: absolute;
top: 0;
left: 0;
width: 100%;
height: 100%;
border: none;
}
</style>
</div>
</div>
<div class="column is-4-desktop is-12-tablet">
<p class="title is-4">Latest News</p>
<div class="columns is-multiline">
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/06/28/MedLog/">MedLog</a>-->
<p class="card-header-title">Jun 2026: <span class="has-text-primary">MedLog</span></p>
<!-- <p class="card-header-item">Jun 2026</p>-->
<!-- <p class="card-footer-item">Jun 28, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>MedLog is an open protocol for event-level logging of medical AI, validated across four real-world pilots in the US, Switzerland, and Vietnam to enable auditing, monitoring, and governance of AI systems. <a href="https://arxiv.org/abs/2510.04033">[Paper]</a> <a href="https://medlogprotocol.ai/">[Project website]</a></p>
</p>-->
<p>MedLog is an open protocol for event-level logging of medical AI, validated across four real-world pilots in the US, Switzerland, and Vietnam to enable auditing, monitoring, and governance of AI systems. <a href="https://arxiv.org/abs/2510.04033">[Paper]</a> <a href="https://medlogprotocol.ai/">[Project website]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/06/28/MedLog/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Jun 28, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/06/22/BiologicalReasoningModels/">Biological Reasoning Models</a>-->
<p class="card-header-title">Jun 2026: <span class="has-text-primary">Biological Reasoning Models</span></p>
<!-- <p class="card-header-item">Jun 2026</p>-->
<!-- <p class="card-footer-item">Jun 22, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Biological reasoning models combine large language models with models of biological data, including DNA, RNA, and proteins. New preprint on training and evaluating <a href="https://arxiv.org/abs/2606.16517">100+ biological reasoning models</a>.</p>
</p>-->
<p>Biological reasoning models combine large language models with models of biological data, including DNA, RNA, and proteins. New preprint on training and evaluating <a href="https://arxiv.org/abs/2606.16517">100+ biological reasoning models</a>.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/06/22/BiologicalReasoningModels/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Jun 22, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/05/28/AutoScientists/">AutoScientists</a>-->
<p class="card-header-title">May 2026: <span class="has-text-primary">AutoScientists</span></p>
<!-- <p class="card-header-item">May 2026</p>-->
<!-- <p class="card-footer-item">May 28, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Introducing <a href="https://autoscientists.openscientist.ai/"><strong>AutoScientists</strong> - self-organizing agent teams for long-running scientific experimentation</a>. <a href="https://arxiv.org/abs/2605.28655">[Paper]</a> <a href="https://github.com/mims-harvard/AutoScientists">[Code]</a></p>
</p>-->
<p>Introducing <a href="https://autoscientists.openscientist.ai/"><strong>AutoScientists</strong> - self-organizing agent teams for long-running scientific experimentation</a>. <a href="https://arxiv.org/abs/2605.28655">[Paper]</a> <a href="https://github.com/mims-harvard/AutoScientists">[Code]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/05/28/AutoScientists/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: May 28, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/05/20/ICML-papers/">Three papers Accepted at ICML 2026</a>-->
<p class="card-header-title">May 2026: <span class="has-text-primary">Three papers Accepted at ICML 2026</span></p>
<!-- <p class="card-header-item">May 2026</p>-->
<!-- <p class="card-footer-item">May 20, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Three papers <a href="https://icml.cc/">accepted to ICML 2026</a>: (1) <a href="https://arxiv.org/abs/2604.12176">Evaluating relational reasoning in LLMs with REL</a>, (2) <a href="https://arxiv.org/abs/2602.18645">Adaptive time series reasoning via segment selection</a>, and (3) <a href="https://arxiv.org/abs/2507.04704">SPATIA: Multimodal model for prediction and generation of spatial cell phenotypes.</a></p>
</p>-->
<p>Three papers <a href="https://icml.cc/">accepted to ICML 2026</a>: (1) <a href="https://arxiv.org/abs/2604.12176">Evaluating relational reasoning in LLMs with REL</a>, (2) <a href="https://arxiv.org/abs/2602.18645">Adaptive time series reasoning via segment selection</a>, and (3) <a href="https://arxiv.org/abs/2507.04704">SPATIA: Multimodal model for prediction and generation of spatial cell phenotypes.</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/05/20/ICML-papers/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: May 20, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/05/01/AgenticAIForScience/">Agentic AI for Science in Nature Methods</a>-->
<p class="card-header-title">May 2026: <span class="has-text-primary">Agentic AI for Science in Nature Methods</span></p>
<!-- <p class="card-header-item">May 2026</p>-->
<!-- <p class="card-footer-item">May 1, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p><a href="https://nature.com/articles/s41592-026-03088-9">Our work on agentic AI for science featured in Nature Methods</a>. This piece features: (1) <a href="https://aiscientist.tools">ToolUniverse — an open platform enabling AI agents to use scientific tools and databases at scale</a>, (2) <a href="https://clawinstitute.aiscientist.tools">ClawInstitute — shared research boards for long-running collaborative discovery where agents co-develop ideas over time</a>, and (3) <a href="https://medea.openscientist.ai">Medea — an omics AI agent for large-scale biological reasoning and analysis</a>.</p>
</p>-->
<p><a href="https://nature.com/articles/s41592-026-03088-9">Our work on agentic AI for science featured in Nature Methods</a>. This piece features: (1) <a href="https://aiscientist.tools">ToolUniverse — an open platform enabling AI agents to use scientific tools and databases at scale</a>, (2) <a href="https://clawinstitute.aiscientist.tools">ClawInstitute — shared research boards for long-running collaborative discovery where agents co-develop ideas over time</a>, and (3) <a href="https://medea.openscientist.ai">Medea — an omics AI agent for large-scale biological reasoning and analysis</a>.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/05/01/AgenticAIForScience/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: May 1, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/04/25/OptimusKG/">OptimusKG: A Modern Knowledge Graph</a>-->
<p class="card-header-title">Apr 2026: <span class="has-text-primary">OptimusKG: A Modern Knowledge Graph</span></p>
<!-- <p class="card-header-item">Apr 2026</p>-->
<!-- <p class="card-footer-item">Apr 25, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p><a href="https://optimuskg.ai/">OptimusKG</a> brings biomedical knowledge into a modern multimodal knowledge graph. It supports graph AI, knowledge-grounded retrieval with large language models, and discovery workflows that generate and evaluate biomedical hypotheses.</p>
</p>-->
<p><a href="https://optimuskg.ai/">OptimusKG</a> brings biomedical knowledge into a modern multimodal knowledge graph. It supports graph AI, knowledge-grounded retrieval with large language models, and discovery workflows that generate and evaluate biomedical hypotheses.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/04/25/OptimusKG/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Apr 25, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/04/23/REL/">Relational Reasoning of LLMs</a>-->
<p class="card-header-title">Apr 2026: <span class="has-text-primary">Relational Reasoning of LLMs</span></p>
<!-- <p class="card-header-item">Apr 2026</p>-->
<!-- <p class="card-footer-item">Apr 23, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>New preprint on <a href="https://arxiv.org/abs/2604.12176">exploring relational reasoning capabilities in LLMs with REL.</a> <a href="https://zitniklab.hms.harvard.edu/REL/">[Project website]</a> <a href="https://kempnerinstitute.harvard.edu/news/new-study-sheds-light-on-the-kinds-of-reasoning-tasks-that-ai-systems-still-struggle-to-do-well/">[Kempner Institute News]</a></p>
</p>-->
<p>New preprint on <a href="https://arxiv.org/abs/2604.12176">exploring relational reasoning capabilities in LLMs with REL.</a> <a href="https://zitniklab.hms.harvard.edu/REL/">[Project website]</a> <a href="https://kempnerinstitute.harvard.edu/news/new-study-sheds-light-on-the-kinds-of-reasoning-tasks-that-ai-systems-still-struggle-to-do-well/">[Kempner Institute News]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/04/23/REL/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Apr 23, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/04/11/ACL/">ARK Accepted at ACL 2026</a>-->
<p class="card-header-title">Apr 2026: <span class="has-text-primary">ARK Accepted at ACL 2026</span></p>
<!-- <p class="card-header-item">Apr 2026</p>-->
<!-- <p class="card-footer-item">Apr 11, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>ARK is an AI agent for <a href="https://arxiv.org/abs/2601.13969">autonomous knowledge graph exploration with adaptive breadth-depth retrieval.</a> <a href="https://github.com/mims-harvard/ark-agent-cli">[ARK Agent CLI]</a></p>
</p>-->
<p>ARK is an AI agent for <a href="https://arxiv.org/abs/2601.13969">autonomous knowledge graph exploration with adaptive breadth-depth retrieval.</a> <a href="https://github.com/mims-harvard/ark-agent-cli">[ARK Agent CLI]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/04/11/ACL/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Apr 11, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/03/21/OpenAIScientists/">Open 'AI Scientists' Initiative</a>-->
<p class="card-header-title">Mar 2026: <span class="has-text-primary">Open 'AI Scientists' Initiative</span></p>
<!-- <p class="card-header-item">Mar 2026</p>-->
<!-- <p class="card-footer-item">Mar 21, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Excited to launch <strong>Open AI Scientists, our initiative to empower scientific discovery with AI scientists.</strong> <a href="https://www.openscientist.ai">[https://www.openscientist.ai]</a></p>
</p>-->
<p>Excited to launch <strong>Open AI Scientists, our initiative to empower scientific discovery with AI scientists.</strong> <a href="https://www.openscientist.ai">[https://www.openscientist.ai]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/03/21/OpenAIScientists/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Mar 21, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/03/20/NatBiotech/">Generalist Biological AI in Nature Biotechnology</a>-->
<p class="card-header-title">Mar 2026: <span class="has-text-primary">Generalist Biological AI in Nature Biotechnology</span></p>
<!-- <p class="card-header-item">Mar 2026</p>-->
<!-- <p class="card-footer-item">Mar 20, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Review on <a href="https://www.nature.com/articles/s41587-026-03064-w">Generalist biological artificial intelligence in modeling the language of life</a> in Nature Biotechnology.</p>
</p>-->
<p>Review on <a href="https://www.nature.com/articles/s41587-026-03064-w">Generalist biological artificial intelligence in modeling the language of life</a> in Nature Biotechnology.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/03/20/NatBiotech/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Mar 20, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/03/16/ClawInstitute/">Claw Institute</a>-->
<p class="card-header-title">Mar 2026: <span class="has-text-primary">Claw Institute</span></p>
<!-- <p class="card-header-item">Mar 2026</p>-->
<!-- <p class="card-footer-item">Mar 16, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p><a href="https://clawinstitute.aiscientist.tools/">Claw Institute is a research exchange for AI scientists</a>. It gives agents a shared space to publish ideas, challenge claims, use scientific tools, and build on one another’s work. These early interactions point to a <a href="https://clawinstitute.aiscientist.tools/blog">new mode of discovery in which societies of AI scientists</a> participate in discovery loops alongside human researchers.</p>
</p>-->
<p><a href="https://clawinstitute.aiscientist.tools/">Claw Institute is a research exchange for AI scientists</a>. It gives agents a shared space to publish ideas, challenge claims, use scientific tools, and build on one another’s work. These early interactions point to a <a href="https://clawinstitute.aiscientist.tools/blog">new mode of discovery in which societies of AI scientists</a> participate in discovery loops alongside human researchers.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/03/16/ClawInstitute/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Mar 16, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/03/07/ICLR-papers/">Five Papers Accepted at ICLR 2026</a>-->
<p class="card-header-title">Mar 2026: <span class="has-text-primary">Five Papers Accepted at ICLR 2026</span></p>
<!-- <p class="card-header-item">Mar 2026</p>-->
<!-- <p class="card-footer-item">Mar 7, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p><a href="https://zitniklab.hms.harvard.edu/publications/">Five papers accepted to ICLR 2026:</a> (1) <a href="https://openreview.net/pdf?id=7LoFonLZqs">Multi-scale protein representation learning</a>, (2) <a href="https://openreview.net/pdf?id=55e5f3GVFc">Protein structure tokenization</a>, (3) <a href="https://openreview.net/pdf?id=gQRefH8upx">Knowledge-based approaches for medical reasoning</a>, (4) <a href="https://openreview.net/pdf?id=dvB0QDmnry">Sequence models for generating biological and health trajectories</a>, and (5) <a href="https://openreview.net/pdf?id=wNEzRYiyZM">Multimodal language models for time series</a>.</p>
</p>-->
<p><a href="https://zitniklab.hms.harvard.edu/publications/">Five papers accepted to ICLR 2026:</a> (1) <a href="https://openreview.net/pdf?id=7LoFonLZqs">Multi-scale protein representation learning</a>, (2) <a href="https://openreview.net/pdf?id=55e5f3GVFc">Protein structure tokenization</a>, (3) <a href="https://openreview.net/pdf?id=gQRefH8upx">Knowledge-based approaches for medical reasoning</a>, (4) <a href="https://openreview.net/pdf?id=dvB0QDmnry">Sequence models for generating biological and health trajectories</a>, and (5) <a href="https://openreview.net/pdf?id=wNEzRYiyZM">Multimodal language models for time series</a>.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/03/07/ICLR-papers/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Mar 7, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/02/23/Overton/">Overton Prize</a>-->
<p class="card-header-title">Feb 2026: <span class="has-text-primary">Overton Prize</span></p>
<!-- <p class="card-header-item">Feb 2026</p>-->
<!-- <p class="card-footer-item">Feb 23, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Our research has been recognized with the <strong>2026 Overton Prize.</strong></p>
</p>-->
<p>Our research has been recognized with the <strong>2026 Overton Prize.</strong></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/02/23/Overton/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Feb 23, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/02/22/StratCP/">Foundation Models that Can 'Act or Defer'</a>-->
<p class="card-header-title">Feb 2026: <span class="has-text-primary">Foundation Models that Can 'Act or Defer'</span></p>
<!-- <p class="card-header-item">Feb 2026</p>-->
<!-- <p class="card-footer-item">Feb 22, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>New preprint on the <a href="https://www.medrxiv.org/content/10.64898/2026.02.23.26346927">safe use of medical foundation models for real-world decision making</a>. <a href="https://zitniklab.hms.harvard.edu/projects/StratCP">[Project website]</a></p>
</p>-->
<p>New preprint on the <a href="https://www.medrxiv.org/content/10.64898/2026.02.23.26346927">safe use of medical foundation models for real-world decision making</a>. <a href="https://zitniklab.hms.harvard.edu/projects/StratCP">[Project website]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/02/22/StratCP/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Feb 22, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/02/22/ARTIST/">Reasoning Model for Longitudinal Data</a>-->
<p class="card-header-title">Feb 2026: <span class="has-text-primary">Reasoning Model for Longitudinal Data</span></p>
<!-- <p class="card-header-item">Feb 2026</p>-->
<!-- <p class="card-footer-item">Feb 22, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>New preprint on <a href="https://arxiv.org/abs/2602.18645">multimodal reasoning model for time series</a>.</p>
</p>-->
<p>New preprint on <a href="https://arxiv.org/abs/2602.18645">multimodal reasoning model for time series</a>.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/02/22/ARTIST/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Feb 22, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/02/15/STRAND/">DNA-Conditioned Models of Single-Cell Perturbations</a>-->
<p class="card-header-title">Feb 2026: <span class="has-text-primary">DNA-Conditioned Models of Single-Cell Perturbations</span></p>
<!-- <p class="card-header-item">Feb 2026</p>-->
<!-- <p class="card-footer-item">Feb 15, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Introducing <a href="https://www.arxiv.org/abs/2602.10156">STRAND, a generative model for predicting how DNA sequence-resolved perturbations shift single-cell gene expression states.</a></p>
</p>-->
<p>Introducing <a href="https://www.arxiv.org/abs/2602.10156">STRAND, a generative model for predicting how DNA sequence-resolved perturbations shift single-cell gene expression states.</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/02/15/STRAND/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Feb 15, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/02/03/Context-Switching/">Context Switching AI in Nature Medicine</a>-->
<p class="card-header-title">Feb 2026: <span class="has-text-primary">Context Switching AI in Nature Medicine</span></p>
<!-- <p class="card-header-item">Feb 2026</p>-->
<!-- <p class="card-footer-item">Feb 3, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>New paper on <a href="https://rdcu.be/e15sO">scaling medical AI across clinical contexts in Nature Medicine</a>.</p>
</p>-->
<p>New paper on <a href="https://rdcu.be/e15sO">scaling medical AI across clinical contexts in Nature Medicine</a>.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/02/03/Context-Switching/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Feb 3, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/01/22/ARK/">Zoom-Out and Zoom-In Retrieval for LLMs</a>-->
<p class="card-header-title">Jan 2026: <span class="has-text-primary">Zoom-Out and Zoom-In Retrieval for LLMs</span></p>
<!-- <p class="card-header-item">Jan 2026</p>-->
<!-- <p class="card-footer-item">Jan 22, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Much of the world’s knowledge lies outside public web text accessible to LLMs, including internal ontologies, curated catalogs, drug safety tables, patient health data, and lab knowledge bases. <a href="https://arxiv.org/abs/2601.13969">ARK helps an LLM to choose, one step at a time, whether to look broadly for relevant information or to dig deeper by following specific links in the data</a>.</p>
</p>-->
<p>Much of the world’s knowledge lies outside public web text accessible to LLMs, including internal ontologies, curated catalogs, drug safety tables, patient health data, and lab knowledge bases. <a href="https://arxiv.org/abs/2601.13969">ARK helps an LLM to choose, one step at a time, whether to look broadly for relevant information or to dig deeper by following specific links in the data</a>.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/01/22/ARK/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Jan 22, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/01/20/Medea/">AI Scientist for Therapeutic Discovery</a>-->
<p class="card-header-title">Jan 2026: <span class="has-text-primary">AI Scientist for Therapeutic Discovery</span></p>
<!-- <p class="card-header-item">Jan 2026</p>-->
<!-- <p class="card-footer-item">Jan 20, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>New preprint on <a href="https://www.biorxiv.org/content/10.64898/2026.01.16.696667">Medea, an omics AI agent for therapeutic discovery.</a> <a href="https://medea.openscientist.ai/">[Project website]</a></p>
</p>-->
<p>New preprint on <a href="https://www.biorxiv.org/content/10.64898/2026.01.16.696667">Medea, an omics AI agent for therapeutic discovery.</a> <a href="https://medea.openscientist.ai/">[Project website]</a></p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/01/20/Medea/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Jan 20, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href="/2026/01/12/ToolUniverse/">AI Scientists - LLMs Using Scientific Tools</a>-->
<p class="card-header-title">Jan 2026: <span class="has-text-primary">AI Scientists - LLMs Using Scientific Tools</span></p>
<!-- <p class="card-header-item">Jan 2026</p>-->
<!-- <p class="card-footer-item">Jan 12, 2026</p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p><p>Excited about this <a href="https://anthropic.com/news/healthcare-life-sciences">academic collaboration with Anthropic</a> on adding connectors to <a href="https://aiscientist.tools/">ToolUniverse</a> to make Claude even more powerful for scientific discovery.</p>
</p>-->
<p>Excited about this <a href="https://anthropic.com/news/healthcare-life-sciences">academic collaboration with Anthropic</a> on adding connectors to <a href="https://aiscientist.tools/">ToolUniverse</a> to make Claude even more powerful for scientific discovery.</p>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="/2026/01/12/ToolUniverse/" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
<!-- <footer class="card-footer">-->
<!-- <p class="card-footer-item">Published: Jan 12, 2026</p>-->
<!-- </footer>-->
</div>
</div>
<div class="column is-12">
<div class="card">
<header class="card-header">
<!-- <a class="card-header-title" href=""></a>-->
<p class="card-header-title"><span class="has-text-primary">Tweets</span></p>
<!-- <p class="card-header-item"></p>-->
<!-- <p class="card-footer-item"></p>-->
</header>
<div class="card-content">
<!-- <div class="content">-->
<!-- -->
<!-- <p></p>-->
<a class="twitter-timeline" data-width="500" data-height="500" data-theme="light" href="https://twitter.com/marinkazitnik?ref_src=twsrc%5Etfw" data-chrome="noheader transparent nofooter nofooter noborders">Tweets by marinkazitnik</a> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>
<!-- </div>-->
<!-- <div class="has-text-centered">-->
<!-- <a href="" class="button is-primary">Read more</a>-->
<!-- </div>-->
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<script src="/assets/js/app.js" type="text/javascript"></script><!-- footer scripts -->
<div style="background-color:#A41034">
<div class="content is-normal has-text-centered">
<p style="color:white;padding-top:20px;padding-bottom:20px;"><a href="https://scholar.harvard.edu/marinka" style="color:white"><b>Zitnik Lab</b></a>
· <a href="#" style="color:white"><b>Artificial Intelligence in Medicine and Science</b></a>
· <a href="https://harvard.edu" style="color:white"><b>Harvard</b></a>
· <a href="https://dbmi.hms.harvard.edu/" style="color:white"><b>Department of Biomedical Informatics</b></a></p>
</div>
</div></body>
</html>