forked from jonbarron/jonbarron.github.io
-
Notifications
You must be signed in to change notification settings - Fork 2
Expand file tree
/
Copy pathindex.html
More file actions
1005 lines (897 loc) · 51.1 KB
/
index.html
File metadata and controls
1005 lines (897 loc) · 51.1 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
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Mengdi Xu</title>
<meta name="author" content="Mengdi Xu">
<meta name="viewport" content="width=900">
<link rel="stylesheet" type="text/css" href="stylesheet.css">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/4.7.0/css/font-awesome.min.css">
</head>
<body>
<div class="navbar">
<div class="container">
<a class="nav-btn" href="#home">Home</a>
<!-- <a class="nav-btn" href="#group">Group</a> -->
<a class="nav-btn" href="#news">News</a>
<a class="nav-btn" href="#publications">Publications</a>
<!-- <a class="nav-btn" href="#talks">Talks</a> -->
<!-- <a class="nav-btn" href="#teaching">Teaching</a> -->
<!-- <a class="nav-btn" href="#awards">Awards</a> -->
<a class="nav-btn" href="#service">Service</a>
<!-- <a class="nav-btn" href="#teaching">Teaching</a> -->
<a class="nav-btn" href="#mentoring">Mentoring</a>
</div>
</div>
<br>
<br>
<section class="content" id="home">
<tr>
<td>
<table>
<tr>
<td width="65%" valign="left">
<p style="text-align:center">
<name>Mengdi Xu</name>
</p>
<p>
Hi! I am an assistant professor at the Institute for Interdisciplinary Information Sciences (IIIS) at Tsinghua University.
Previously, I was a postdoc at <a href="https://svl.stanford.edu/">Stanford Vision and Learning Lab (SVL)</a> working with Prof. <a href="https://jiajunwu.com/">Jiajun Wu</a> and Prof. <a href="https://profiles.stanford.edu/fei-fei-li">Fei-Fei Li</a>. I received my PhD from Carnegie Mellon University with Best PhD Dissertation Award, advised by Prof. <a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html">Ding Zhao</a>
(<a href="https://safeai-lab.github.io/">SafeAI Lab</a>). My thesis is about building adaptable generalist robots.
I have also spent wonderful summers at <a href="https://blog.google/technology/ai/april-ai-update/">Google DeepMind</a> Robotics, <a href="https://mitibmwatsonailab.mit.edu/">MIT-IBM Watson AI Lab</a>, and <a href="https://www.tri.global/">Toyota Research Institute</a>.
<!-- Previously, I received a master's degree in Machine Learning from <a href="https://www.ml.cmu.edu/">Machine Learning Department</a> at CMU, and a master's degree in Robotics from <a href="https://lcsr.jhu.edu/">LCSR</a> at JHU, supervised by Prof. <a href="https://rpk.lcsr.jhu.edu/Publications/">Gregory S. Chirikjian</a>. -->
I received a B.E. from Tsinghua University.
<br>
<br>
<p style="text-align:center">
<a href="mailto:xumd0105@gmail.com">Email</a>  / 
<a href="data/CV_MengdiXu.pdf">CV (Mar 2024)</a>  / 
<a href="https://scholar.google.com/citations?hl=zh-CN&user=A1yMyYEAAAAJ">Google Scholar</a>  / 
<a href="https://github.com/mxu34">Github</a>  / 
<a href="https://x.com/mengdixu_">X</a>
</p>
</p>
<br>
<p> <strong>I am actively recruiting students to join my lab at Tsinghua IIIS and am also open to research collaborations. If you are interested, please feel free to reach out! </strong></p>
</td>
<br>
<td width="25%" valign="right" style="padding-left: 30px;">
<a href="images/profile_24.png"><img style="width:100%;max-width:100%;border-radius:8px" alt="profile photo" src="images/profile_24.jpg" class="hoverZoomLink"></a>
</td>
</tr>
</table>
</td>
</tr>
<br>
<details>
<summary><strong>Prospective Students & Collaborations</strong></summary>
<p>
Thank you for your interest in our lab! We welcome highly motivated undergraduates, graduate students, and visiting scholars to join us and collaborate. We are a highly interdisciplinary lab seeking students from diverse backgrounds, including robot hardware design, large-scale model training, machine learning, control, and robotics.
<br>
For undergraduates and graduate students, please complete <a href="https://dcnpy29oktbo.feishu.cn/share/base/form/shrcnTa8Qmkn6ZdX4ltWiwsyEBg">this form</a> and send me an email.
<br>
For potential collaborators, feel free to email me directly with the projects you are interested in.
</p>
</details>
</section>
<section class="content">
<tr>
<td>
<h2> Research </h2>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<p>
I am broadly interested in building scalable, adaptable, and reliable robots that seamlessly interact with humans in daily activities. Here are some research highlights:
<ul>
<li><strong>In-context robot learning</strong> with few-shot demonstrations and non-expert data.</li>
<li><strong>Unsupervised learning</strong> to discover task/skill structures in dynamics, visual geometries, and agent policies.</li>
<li><strong>Distributionally robust RL</strong> that balances performance and robustness when facing task uncertainties.</li>
<!-- <li>laying theoretical foundations for generalizable robot learning with statistical guarantees,</li> -->
<!-- <li>advancing robotics with novel capabilities in applications akin to everyday life.</li> -->
</ul>
</p>
</td>
</tr>
</table>
</td>
</tr>
</section>
<section class="content" id="group">
<h2> Group </h2>
<p>
PhD students:
<!-- <ul>
<li>Liu Cao (2025-)</li>
<li>Chenyu Zhang (2025-)</li>
<li>Ruoqu Chen (2026-)</li>
</ul> -->
</p>
<div class="four-photos">
<figure>
<img src="group/liu.jpeg" alt="photo 1">
<figcaption><a href="https://xiaohu-art.github.io/">Liu Cao</a> (2025-)</figcaption>
</figure>
<figure>
<img src="group/chenyu.jpeg" alt="photo 1">
<figcaption>Chenyu Zhang (2025-)</figcaption>
</figure>
<figure>
<img src="group/ruoqu.jpg" alt="photo 1">
<figcaption><a href="https://gcfy63821.github.io">Ruoqu Chen</a> (2026-)</figcaption>
</figure>
</div>
</section>
<section class="content" id="news">
<tr>
<td>
<h2> News </h2>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<p>
2026/02 - We released <a href="https://cover-vla.github.io/">CoVer-VLA</a>, a test-time scaling framework for VLA.
<br>
2026/02 - Honored to serve as a Program Committee (PC) member for RSS Pioneers 2026.
<br>
2026/01 - <a href="https://momagen.github.io/">MoMaGen</a> and <a href="https://sanjanasrivastava.github.io/rosetta-project/">ROSETTA</a> got accepted to ICLR 2026!
<br>
2025/10 - <a href="https://momagen.github.io/">MoMaGen</a> released. Check out how MoMaGen generates data for long-horizon bimanual mobile manipulation tasks.
<br>
2025/09 - I joined Tsinghua IIIS as a tenure-track Assistant Professor.
<br>
2025/06 - ROSETTA won the <strong style="color:#B908C5">best paper award</strong> at the RSS 2025 Workshop on Continual Robot Learning from Humans (<a href="https://continual-robot-learning-from-humans.github.io/">CRLH</a>).
<br>
2024/08 - Functional Optimal Transport got accepted to JMLR.
<br>
2024/07 - Invited talk at <a href="https://lecar-lab.github.io/">CMU LeCAR Lab</a>.
<br>
2024/06 - Joined <a href="https://svl.stanford.edu/">Stanford Vision and Learning Lab (SVL)</a> as a postdoctoral researcher.
<br>
2024/05 - We are organizing the RSS Workshop on <a href="https://sites.google.com/view/lifelong-robot-learning">Lifelong Robot Learning: Generalization, Adaptation, and Deployment with Large Models</a> and <a href="https://sites.google.com/view/rsspioneers2024/">RSS Pioneers 2024 Workshop</a>. See you in Delft!
<br>
<!-- 2024/05 - I successfully passed my thesis defense. Many thanks to my thesis committee!
<br> -->
<!-- 2024/03 - One paper got accepted to NAACL 2024.
<br> -->
<!-- 2023/10 - A new arXiv preprint <a href="https://arxiv.org/pdf/2310.13065.pdf">RoboTool</a>.
<br> -->
<!-- 2023/09 - One paper got accepted to NeurIPS 2023.
<br> -->
<!-- 2023/08 - Two papers got accepted to CoRL 2023, and <a href="https://openreview.net/forum?id=flyQ0v8cgC">COVERS</a> was accepted for an oral presentation.
<br>
2023/08 - Selected as <a href="https://eecsrisingstars2023.cc.gatech.edu/">EECS Rising Star</a>.
<br>
2023/05 - Selected as <a href="https://sites.google.com/view/rsspioneers2023/participants">RSS Pioneer</a>.
<br>
2023/02 - Selected as <a href="https://risingstars.oden.utexas.edu/">Rising Star in Computational and Data Sciences</a>.
<br> -->
<!-- 2023/02 - I joined Google DeepMind as a Student Researcher. -->
<!-- <br> -->
<!-- 2023/01 - Our work about bridging ECG and NLP got accepted to EACL 2023.
<br> -->
<!-- 2023/01 - Our work on efficient adaptation of transformer agents got accepted to ICLR 2023. -->
<!-- <br>
2023/01 - Our work about group distributionally robust RL got accepted to AISTATS 2023.
<br>
2022/09 - Our work about domain adaptation in curriculum RL got accepted to NuerIPS 2022.
<br>
2022/06 - Our work on accelerated policy evaluation in the presence of rare events got accepted to IROS 2022.
<br>
2022/05 - I joined MIT-IBM Watson AI Lab as a research intern.
<br>
2022/05 - Our work on policy generalization with offline data got accepted to ICML 2022.
<br>
2022/04 - Our work about a novel framework for robust RL got accepted to IJCAI 2022.
<br> -->
</p>
</td>
</tr>
</table>
</td>
</tr>
</section>
<section class="content" id="publications">
<tr>
<td>
<h2> Selected Publications </h2>
<p>
A full list of publications is <a href="full_pub.html">here</a>. (* indicates equal contribution.)
</p>
<h3>Data Generation</h3>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/momagen_method.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>MoMaGen: Generating Demonstrations under Soft and Hard Constraints for Multi-Step Bimanual Mobile Manipulation
</papertitle>
<br>
<a href="https://www.chengshuli.me/"><author>Chengshu Li*</author></a>,
<strong>Mengdi Xu*</strong>,
<a href="https://arpitrf.github.io/"><author>Arpit Bahety*</author></a>,
<a href="https://hang-yin.github.io/"><author>Hang Yin*</author></a>,
<a href="https://yunfanj.com/"><author>Yunfan Jiang</author></a>,
<a href="https://qingh097.github.io/"><author>Huang Huang</author></a>,
<a href="https://www.jowo.me/"><author>Josiah Wong</author></a>,
<a href=""><author>Sujay Garlanka</author></a>,
<a href="https://www.cemgokmen.com/"><author>Cem Gokmen</author></a>,
<a href="https://ruohanzhang.com/"><author>Ruohan Zhang</author></a>,
<a href="https://www.weiyuliu.com/"><author>Weiyu Liu</author></a>,
<a href="https://jiajunwu.com/"><author>Jiajun Wu</author></a>,
<a href="https://robertomartinmartin.com/"><author>Roberto Martín-Martín</author></a>,
<a href="https://profiles.stanford.edu/fei-fei-li"><author>Fei-Fei Li</author></a>
<br>
<strong>ICLR</strong> 2026.
<br>
<a href="https://www.arxiv.org/abs/2510.18316">[paper]</a>
<a href="https://momagen.github.io/">[website]</a>
<a href="https://github.com/ChengshuLi/MoMaGen">[code]</a>
</papertext>
</td>
</tr>
</table>
<br>
<h3>Few-shot Generalization</h3>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/cover.gif" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Scaling Verification Can Be More Effective than Scaling Policy Learning for Vision-Language-Action Alignment
</papertitle>
<br>
<a href="https://www.linkedin.com/in/jackykwok02/"><author>Jacky Kwok*</author></a>,
<a href="https://www.cs.cmu.edu/~jielinq/"><author>Xilun Zhang*</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://sites.google.com/view/yuejiangliu/home"><author>Yuejiang Liu†</author></a>,
<a href="https://www.azaliamirhoseini.com/"><author>Azalia Mirhoseini†</author></a>,
<a href="https://ai.stanford.edu/~cbfinn/"><author>Chelsea Finn†</author></a>,
<a href="https://research.nvidia.com/person/marco-pavone"><author>Marco Pavone†</author></a>
<br>
In submission, 2026. † equal advising.
<br>
<a href="https://arxiv.org/abs/2602.12281">[paper]</a>
<a href="https://cover-vla.github.io/">[website]</a>
</papertext>
</td>
</tr>
</table>
<br>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/robotool.gif" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Creative Robot Tool Use with Large Language Models</papertitle>
<br>
<strong>Mengdi Xu*</strong>,
<a href="https://peidehuang.github.io/"><author>Peide Huang*</author></a>,
<a href="https://wenhaoyu.weebly.com/"><author>Wenhao Yu*</author></a>,
<a href="https://scholar.google.com/citations?user=PiuAi5wAAAAJ&hl=en&oi=ao"><author>Shiqi Liu</author></a>,
<a href="https://scholar.google.com/citations?user=BnlIUZoAAAAJ&hl=en"><author>Xilun Zhang</author></a>,
<a href="https://yaruniu.com/"><author>Yaru Niu</author></a>,
<a href="https://scholar.google.com/citations?user=RM2vMNcAAAAJ&hl=en"><author>Tingnan Zhang</author></a>,
<a href="https://fxia22.github.io/"><author>Fei Xia</author></a>,
<a href="https://www.jie-tan.net/"><author>Jie Tan</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
CoRL 2023 Workshop on Language and Robot Learning: Language as Grounding
<br>
<a href="https://arxiv.org/pdf/2310.13065.pdf">[paper]</a>
<a href="https://creative-robotool.github.io/">[webpage]</a>
<a href="https://blog.ml.cmu.edu/2023/12/08/creative-robot-tool-use-with-large-language-models/">[MLD Blog]</a>
<a href="https://techxplore.com/news/2023-11-robots-tools-creatively-leveraging-large.html">[TechXplore]</a>
</papertext>
</td>
</tr>
</table>
<br>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/HDT.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Hyper-Decision Transformer for Efficient Online Policy Adaptation</papertitle>
<br>
<strong>Mengdi Xu</strong>,
<a href="http://jackhaha363.github.io/"><author>Yuchen Lu</author></a>,
<a href="https://mitibmwatsonailab.mit.edu/people/yikang-shen/"><author>Yikang Shen</author></a>,
<a href="https://shunzh.github.io/"><author>Shun Zhang</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>,
<a href="http://people.csail.mit.edu/ganchuang/"><author>Chuang Gan</author></a>
<br>
The Eleventh International Conference on Learning Representations (<strong>ICLR</strong>), 2023
<br>
<a href="https://arxiv.org/pdf/2304.08487.pdf">[paper]</a>
<a href="https://sites.google.com/view/hdtforiclr2023/home">[webpage]</a>
</papertext>
</td>
</tr>
</table>
<br>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/Prompt_DT_website.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Prompting Decision Transformer for Few-shot Policy Generalization
</papertitle>
<br>
<strong>Mengdi Xu</strong>,
<a href="https://mitibmwatsonailab.mit.edu/people/yikang-shen/"><author>Yikang Shen</author></a>,
<a href="https://shunzh.github.io/"><author>Shun Zhang</author></a>,
<a href="http://jackhaha363.github.io/"><author>Yuchen Lu</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>,
<a href="http://web.mit.edu/cocosci/josh.html"><author>Josh Tenenbaum</author></a>,
<a href="http://people.csail.mit.edu/ganchuang/"><author>Chuang Gan</author></a>
<br>
Thirty-ninth International Conference on Machine Learning (<strong>ICML</strong>), 2022
<br>
<a href="https://arxiv.org/pdf/2206.13499.pdf">[paper]</a>
<a href="https://mxu34.github.io/PromptDT/">[webpage]</a>
<a href="https://github.com/mxu34/prompt-dt">[code]</a>
</papertext>
</td>
</tr>
</table>
<br>
<h3>Efficient Adaptation</h3>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/rosetta.jpg" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>ROSETTA: Constructing Code-Based Reward from Unconstrained Language Preference
</papertitle>
<br>
<a href="https://www.linkedin.com/in/sanjana-srivastava5/"><author>Sanjana Srivastava*</author></a>,
<a href="https://jameskrw.github.io/"><author>Kangrui Wang*</author></a>,
<a href="https://github.com/JerryYC"><author>Yung-Chieh Chan*</author></a>,
<a href="https://rogerdai1217.github.io/"><author>Tianyuan Dai</author></a>,
<a href="https://limanling.github.io/"><author>Manling Li</author></a>,
<a href="https://ruohanzhang.com/"><author>Ruohan Zhang</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://jiajunwu.com/"><author>Jiajun Wu</author></a>,
<a href="https://profiles.stanford.edu/fei-fei-li"><author>Fei-Fei Li</author></a>
<br>
<strong>ICLR</strong> 2026.
<br>
<a href="https://sanjanasrivastava.github.io/rosetta-project/">[website]</a>
</papertext>
</td>
</tr>
</table>
<br>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/covers.gif" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Continual Vision-based Reinforcement Learning with Group Symmetries</papertitle>
<br>
<a href="https://scholar.google.com/citations?user=PiuAi5wAAAAJ&hl=en&oi=ao"><author>Shiqi Liu*</author></a>,
<strong>Mengdi Xu*</strong>,
<a href="https://peidehuang.github.io/"><author>Peide Huang</author></a>,
<a href="https://scholar.google.com/citations?user=uRvYtOgAAAAJ&hl=en"><author>Yongkang Liu</author></a>,
<a href="https://www.linkedin.com/in/ken-oguchi-a89b263/"><author>Kentaro Oguchi</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
Conference on Robot Learning (<strong>CoRL</strong>), 2023 <strong style="color:#B908C5">(oral)</strong>
<br>
<!-- Abridged in RSS 2023 Workshop on Symmetries in Robot Learning.<br> -->
<a href="https://arxiv.org/pdf/2210.12301.pdf">[paper]</a>
<a href="https://sites.google.com/view/rl-covers/">[webpage]</a>
<a href="https://sites.google.com/view/rl-covers/">[code]</a>
</papertext>
</td>
</tr>
</table>
<br>
<!-- <table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/mbrl-gpmm.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Task-Agnostic Online Reinforcement Learning with an Infinite Mixture of Gaussian Processes
</papertitle>
<br>
<strong>Mengdi Xu</strong>,
<a href="https://wenhao.pub/"><author>Wenhao Ding</author></a>,
<a href="https://jiachengzhuml.github.io/"><author>Jiacheng Zhu</author></a>,
<a href="https://zuxin.me/"><author>Zuxin Liu</author></a>,
<a href="https://scholar.google.com/citations?user=zxMrNxEAAAAJ&hl=en&oi=ao"><author>Baiming Chen</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
Thirty-fourth Conference on Neural Information Processing Systems (<strong>NeurIPS</strong>), 2020<br>
<a href="https://arxiv.org/pdf/2006.11441.pdf">[paper]</a>
<a href="https://github.com/mxu34/mbrl-gpmm">[code]</a>
</papertext>
</td>
</tr>
</table> -->
<!-- <table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/compass.gif" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>What Went Wrong? Closing the Sim-to-Real Gap via Differentiable Causal Discovery</papertitle>
<br>
<a href="https://peidehuang.github.io/"><author>Peide Huang</author></a>,
<a href="https://scholar.google.com/citations?user=BnlIUZoAAAAJ&hl=en"><author>Xilun Zhang*</author></a>,
Ziang Cao*,
<a href="https://scholar.google.com/citations?user=PiuAi5wAAAAJ&hl=en&oi=ao"><author>Shiqi Liu*</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://wenhao.pub/"><author>Wenhao Ding</author></a>,
<a href="https://scholar.google.com/citations?user=7CLS0LwAAAAJ&hl=en"><author>Jonathan Francis</author></a>,
<a href="https://scholar.google.com/citations?user=LYt_2MgAAAAJ&hl=en"><author>Bingqing Chen</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
Conference on Robot Learning (<strong>CoRL</strong>), 2023
<br>
<a href="https://arxiv.org/pdf/2306.15864.pdf">[paper]</a>
<a href="https://sites.google.com/view/sim2real-compass">[webpage]</a>
</papertext>
</td>
</tr>
</table>
<br> -->
<!--
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/gradient.gif" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Curriculum Reinforcement Learning using Optimal Transport via Gradual Domain Adaptation
</papertitle>
<br>
<a href="https://peidehuang.github.io/"><author>Peide Huang</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://jiachengzhuml.github.io/"><author>Jiacheng Zhu</author></a>,
<a href="https://www.andrew.cmu.edu/user/laixis/"><author>Laixi Shi</author></a>,
<a href="https://feifang.info/"><author>Fei Fang</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>,
<br>
The 36th Conference on Neural Information Processing Systems, (<strong>NeurIPS</strong>), 2022
<br>
<a href="https://arxiv.org/pdf/2210.10195.pdf">[paper]</a>
<a href="https://github.com/PeideHuang/gradient">[code]</a>
</papertext>
</td>
</tr>
</table> -->
<!-- <table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/SUM_APM.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Embodied Executable Policy Learning with Language-based Scene Summarization
</papertitle>
<br>
<a href="https://www.cs.cmu.edu/~jielinq/"><author>Jielin Qiu</author>*</a>,
<strong>Mengdi Xu*</strong>,
<a href="https://willxxy.github.io/"><author>William Han</author>*</a>,
<a href="https://shanemoon.com/"><author>Seungwhan Moon</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL)
<br>
ICML 2023 Workshop on Interactive Learning with Implicit Human Feedback <strong style="color:#B908C5">(spotlight)</strong><br>
<a href="https://arxiv.org/pdf/2306.05696.pdf">[paper]</a>
</papertext>
</td>
</tr>
</table>
<br> -->
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/FOT_arm.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Functional Optimal Transport: Map Estimation and Domain Adaptation for Functional data
</papertitle>
<br>
<a href="https://jiachengzhuml.github.io/"><author>Jiacheng Zhu*</author></a>,
<a href="http://www-personal.umich.edu/~aritra/"><author>Aritra Guha*</author></a>,
<a href="https://dodat619.github.io/"><author>Dat Do*</author></a>,
<strong>Mengdi Xu</strong>,
<a href="http://dept.stat.lsa.umich.edu/~xuanlong/"><author>XuanLong Nguyen</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
JMLR: Journal of Machine Learning Research
<br>
AAAI OT-SDM 2022 workshop <strong style="color:#B908C5">(spotlight)</strong><br>
<a href="https://arxiv.org/pdf/2102.03895.pdf">[paper]</a>
<a href="https://github.com/VersElectronics/FOT">[code]</a>
</papertext>
</td>
</tr>
</table>
<br>
<h3>Robust and Safe Robot Learning</h3>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/DRO.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Group Distributionally Robust Reinforcement Learning with Hierarchical Latent Variables
</papertitle>
<br>
<strong>Mengdi Xu</strong>,
<a href="https://peidehuang.github.io/"><author>Peide Huang</author></a>,
<a href="https://yaruniu.com/"><author>Yaru Niu</author></a>,
<a href="https://www.visakkumar.com/"><author>Visak Kumar</author></a>,
<a href="https://www.cs.cmu.edu/~jielinq/"><author>Jielin Qiu</author></a>,
<a href="https://scholar.google.com/citations?user=nU8VhKYAAAAJ&hl=en"><author>Chao Fang</author></a>,
<a href="https://scholar.google.com/citations?user=u8K7nOwAAAAJ&hl=en"><author>Kuan-Hui Lee</author></a>,
<a href="https://scholar.google.com/citations?user=pOA6uKMAAAAJ&hl=en"><author>Xuewei Qi</author></a>,
<a href="http://www.columbia.edu/~khl2114/"><author>Henry Lam</author></a>,
<a href="https://aisecure.github.io/"><author>Bo Li</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
The 26th International Conference on Artificial Intelligence and Statistics, (<strong>AISTATS</strong>), 2023
<!-- <br>
Abridged in ICML 2022 workshop on Principles of Distribution Shift.
<br>
Abridged in ICRA 2022 the Fresh Perspectives on the Future of Autonomous Driving Workshop. -->
<br>
<a href="https://arxiv.org/pdf/2210.12262.pdf">[paper]</a>
<a href="https://sites.google.com/view/gdr-rl/home">[webpage]</a>
</papertext>
</td>
</tr>
</table>
<br>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/APE.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Scalable Safety-Critical Policy Evaluation with Accelerated Rare Event Sampling
</papertitle>
<br>
<strong>Mengdi Xu</strong>,
<a href="https://peidehuang.github.io/"><author>Peide Huang</author></a>,
<a href="https://www.linkedin.com/in/fengpei-li-242526145/"><author>Fengpei Li</author></a>,
<a href="https://jiachengzhuml.github.io/"><author>Jiacheng Zhu</author></a>,
<a href="https://scholar.google.com/citations?user=pOA6uKMAAAAJ&hl=en"><author>Xuewei Qi</author></a>,
<a href="https://www.linkedin.com/in/ken-oguchi-a89b263/"><author>Kentaro Oguchi</author></a>,
<a href="https://sites.google.com/view/zyhuang"><author>Zhiyuan Huang</author></a>,
<a href="http://www.columbia.edu/~khl2114/"><author>Henry Lam</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
IEEE/RSJ International Conference on Intelligent Robots and Systems (<strong>IROS</strong>), 2022
<br>
<!-- Abridged in ICLR 2021 Workshop in Security and Safety in Machine Learning Systems.
<br> -->
<a href="https://arxiv.org/pdf/2106.10566.pdf">[paper]</a>
<!-- <a href="data/ape.txt">[bibtex]</a> -->
<!-- <a href="https://aisecure-workshop.github.io/aml-iclr2021/">[workshop]</a>
<a href="data/APE_ICLR2021Workshop_poster.pdf">[poster]</a> -->
</papertext>
</td>
</tr>
</table>
<br>
<!-- <table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/rrl-stack.gif" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Robust Reinforcement Learning as a Stackelberg Game via Adaptively-Regularized Adversarial Training
</papertitle>
<br>
<a href="https://peidehuang.github.io/"><author>Peide Huang</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://feifang.info/"><author>Fei Fang</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
International Joint Conference on Artificial Intelligence (<strong>IJCAI</strong>), 2022
<br>
<a href="https://arxiv.org/pdf/2202.09514.pdf">[paper]</a>
</papertext>
</td>
</tr>
</table>
-->
<!-- <table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/context_RL.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Context-Aware Safe Reinforcement Learning for Non-Stationary Environments
</papertitle>
<br>
<a href="https://scholar.google.com/citations?user=zxMrNxEAAAAJ&hl=en&oi=ao"><author>Baiming Chen</author></a>,
<a href="https://zuxin.me/"><author>Zuxin Liu</author></a>,
<a href="https://jiachengzhuml.github.io/"><author>Jiacheng Zhu</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://wenhao.pub/"><author>Wenhao Ding</author></a>,
Liang Li,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
IEEE International Conference on Robotics and Automation (<strong>ICRA</strong>), 2021<br>
<a href="https://arxiv.org/pdf/2101.00531.pdf">[paper]</a>
</papertext>
</td>
</tr>
</table>
-->
<!-- <table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/dmdpn.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Delay-Aware Model-Based Reinforcement Learning for Continuous Control
</papertitle>
<br>
<a href="https://scholar.google.com/citations?user=zxMrNxEAAAAJ&hl=en&oi=ao"><author>Baiming Chen</author></a>,
<strong>Mengdi Xu</strong>,
Liang Li,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
<strong>Neurocomputing</strong>, 2021<br>
<a href="https://arxiv.org/pdf/2005.05440.pdf">[paper]</a>
<a href="https://github.com/baimingc/Delay-aware-MBRL">[code]</a>
</papertext>
</td>
</tr>
</table>
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/CMTS.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>CMTS: Conditional Multiple Trajectory Synthesizer for Generating Safety-critical Driving Scenarios
</papertitle>
<br>
<a href="https://wenhao.pub/"><author>Wenhao Ding</author></a>,
<strong>Mengdi Xu</strong>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
International Conference on Robotics and Automation (<strong>ICRA</strong>), 2020<br>
<a href="https://arxiv.org/pdf/1910.00099.pdf">[paper]</a>
<a href="https://github.com/GilgameshD/ICRA2020-CMTS">[code]</a>
</papertext>
</td>
</tr>
</table> -->
<table width="950" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 20px;">
<div class="image-container">
<img src="images/TRL.png" width="100%">
</div>
</td>
<td style="width:75%; vertical-align:middle">
<papertext>
<papertitle>Trustworthy Reinforcement Learning Against Intrinsic Vulnerabilities: Robustness, Safety, and Generalizability
</papertitle>
<br>
<strong>Mengdi Xu*</strong>,
<a href="https://zuxin.me/"><author>Zuxin Liu</author>*</a>,
<a href="https://peidehuang.github.io/"><author>Peide Huang</author>*</a>,
<a href="https://wenhao.pub/"><author>Wenhao Ding</author></a>,
<a href="https://scholar.google.com/citations?user=M-X3Q-UAAAAJ&hl=en&oi=ao"><author>Zhepeng Cen</author></a>,
<a href="https://aisecure.github.io/"><author>Bo Li</author></a>,
<a href="https://www.meche.engineering.cmu.edu/directory/bios/zhao-ding.html"><author>Ding Zhao</author></a>
<br>
Preprint, under review
<br>
<a href="https://arxiv.org/pdf/2209.08025.pdf">[paper]</a>
</papertext>
</td>
</tr>
</table>
</td>
</tr>
</section>
<!-- <table width="950" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td>
<h2> Talks </h2>
<hr>
<talktitle>Generalizable Robot Learning: Adaptability, Robustness and Evaluation</talktitle>
<ul>
<li>ByteDance AI Lab, November 2023 </li>
<li>Microsoft Research Lab - Asia, November 2023 </li>
<li>EECS Rising Stars workshop, November 2023 </li>
<li>Hybrid Robotics Lab at Berkeley, hosted by Prof. <a href="https://hybrid-robotics.berkeley.edu/koushil/">Koushil Sreenath</a>, November 2023 </li>
<li>Robotics: Science and Systems (RSS) Pioneers, July 2023 </li>
<li>UT Austin Computational and Data Science Rising Stars workshop, April 2023 </li>
<li>CMU Manipulation Reading Group, April 2023.</li>
</ul>
<talktitle>Continual Vision-based Reinforcement Learning with Group Symmetries</talktitle>
<ul>
<li>CoRL oral presentation, November 2023 </li>
</ul>
<talktitle>Prompting Decision Transformer for Few-shot Policy Generalization</talktitle>
<ul>
<li>AISOC Lab at CMU, hosted by Prof. <a href="https://feifang.info/">Fei Fang</a>, April 2022 </li>
</ul>
<talktitle>Task-Agnostic Online Reinforcement Learning with an Infinite Mixture of Gaussian Processes </talktitle>
<ul>
<li>Hosted by Prof. <a href="https://dept.stat.lsa.umich.edu/~xuanlong/">Long Nguyen</a> at the University of Michigan, November 2020 </li>
</ul>
</td>
</tr>
</table>
<br>
<br> -->
<!-- <section class="content" id="teaching">
<tr>
<td>
<h2> Teaching </h2>
<p>Planned Courses</p>
<ul>
<li>Fall 2025. Interactive AI, Tsinghua University.</li>
<li>Spring 2026. AI Safety, Tsinghua University.</li>
</ul>
</td>
</tr>
</section> -->
<section class="content" id="talks">
<tr>
<td>
<h2>Talks</h2>
<p>Research Overview March 2024</p>
<iframe width="560" height="315"
src="https://www.youtube.com/embed/InklIGD8lvo"
title="YouTube video player"
frameborder="0"
allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share"
allowfullscreen>
</iframe>
</td>
</tr>
</section>
<section class="content" id="awards">
<tr>
<td>
<h2> Awards </h2>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<p>
<ul>
<li>Best PhD Dissertation Award 2024, Mechanical Engineering, CMU, 2024</li>
<li>EECS Rising Stars, 2023</li>
<li>Robotics: Science and Systems (RSS) Pioneers, 2023</li>
<li>Rising Stars in Computational and Data Sciences, 2023</li>
</ul>
</p>
</td>
</tr>
</table>
</td>
</tr>
</section>
<section class="content" id="service">
<tr>
<td>
<h2> Service </h2>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<p>
<ul>
<li>Co-organized RSS 2024 Workshop on Lifelong Robot Learning: Generalization, Adaptation, and Deployment with Large Models</li>
<li>Co-organized RSS Pioneers Workshop 2024</li>
<li>Breakout session leader at 3rd Women in Machine Learning Un-Workshop, ICML 2022</li>
<li>Mentor of 2020 CMU Robotics Institute Summer Scholars Program (RISS)</li>
<li>Conference Reviewer: NeurIPS, ICML, ICLR, CoRL, AISTATS, ACL, ICRA, ICCV, ECCV, CVPR, AAAI, L4DC, ICASSP</li>
<li>Journal Reviewer: IJRR, RA-L, T-ITS, T-IV</li>
</ul>
</p>
</td>
</tr>
</table>
</td>
</tr>
</section>
<!-- <section class="content" id="teaching">
<h2> Teaching </h2>
<p>
<ul>
<li>Interactive AI, 2025 Fall</li>
</ul>
</p>
</section> -->
<section class="content" id="mentoring">
<h2> Mentoring </h2>
<p>
<ul>
<li><a href="https://xilunzhangrobo.github.io/">Xilun Zhang</a> (Master, CMU → PhD, Stanford)</li>
<li><a href="https://shiqiliu-67.github.io/">Shiqi Liu</a> (Master, CMU → PhD, CMU)</li>
<li><a href="https://zi-ang-cao.github.io/">Ziang Cao</a> (Undergrad, UPitt → Master, Stanford) </li>
</ul>
</p>
</section>
<br>
<br>
<br>
<br>
<table width="950" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td>
<p align="right">
Website template from <a href="https://github.com/jonbarron/website">Jon Barron</a>.
</p>
</td>
</tr>
</table>