-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
1728 lines (1715 loc) · 73.1 KB
/
Copy pathindex.html
File metadata and controls
1728 lines (1715 loc) · 73.1 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
<!---
remember to replace – into -
--->
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<link rel="stylesheet" href="styles.css">
</head>
<title>
Shu-Jian HUANG’s HomePage
</title>
<p><span id="top"></p>
<p><img src="resources/head.jpg" /></p>
<p><img src="resources/signature.jpg" /></p>
<h1 id="shu-jian-huang-ph.d.">Shu-Jian Huang Ph.D.</h1>
<p>Professor, Ph. D. Advisor</p>
<p><a href="https://nlp.nju.edu.cn/">Natural Language Processing
Research Group</a></p>
<p>National Key Laboratory of Novel Software Technology</p>
<p><a href="https://cs.nju.edu.cn/">School of Computer Science</a>,
Nanjing University</p>
<hr />
<p><a href="#correspondence">[Correspondence]</a> <a
href="#biography">[Biography]</a> <a href="#recruiting">[Recruiting]</a>
<a href="#interests">[Research Interests]</a> <a
href="#services">[Academic Services]</a> <a href="#honors">[Honors and
Awards]</a> <a href="#courses">[Courses]</a> <a
href="#presentations">[Presentations]</a> <a href="#pub">[Selected
Publication]</a></p>
<hr />
<p><span id="correspondence"></p>
<h2 id="correspondence">Correspondence</h2>
<p>Office:</p>
<blockquote>
<p>Room 902, Computer Science and Technology Building, Xianlin Campus of
Nanjing University</p>
</blockquote>
<blockquote>
<p>163 Xianlin Avenue, Nanjing 210023, China</p>
</blockquote>
<p>E-mail:</p>
<blockquote>
<p>huangsj at nju dot edu dot cn</p>
</blockquote>
<p>Homepage:</p>
<blockquote>
<p><a
href="https://nlp.nju.edu.cn/huangsj">https://nlp.nju.edu.cn/huangsj</a></p>
</blockquote>
<blockquote>
<p><a href="https://shujian.github.io">https://shujian.github.io</a></p>
</blockquote>
<hr />
<p><span id="biography"></p>
<h2 id="biography">Biography</h2>
<blockquote>
<p>Currently, I am a professor and Ph. D. Advisor in School of Computer
Science (formerly known as the Department of Computer Science and
Technology) of Nanjing University, member of the State Key Laboratory of
Novel Software Technology.</p>
</blockquote>
<blockquote>
<p>I received my B.Sc. degree and Ph. D. in Computer Science in
Jun. 2006 and Jun. 2012 from Nanjing University, respectively. I am a
member of <a href="https://nlp.nju.edu.cn/">NLP Group</a> since
undergraduate in Sep. 2005, led by <a
href="https://cs.nju.edu.cn/chenjiajun/index_en.htm">Prof. Jiajun
Chen</a>. During my Ph. D. study, I spent 11 months (from Oct. 2007 to
Aug. 2008) as visiting student in <a
href="https://research.microsoft.com/en-us/groups/nlc/">NLC group</a>,
MSRA, where I worked with Long Jiang in <a
href="https://couplet.msra.cn/">Chinese Couplet project</a> and in SMT
team with Mu Li, Henry Li and Dongdong Zhang. I also spent 12 months as
a visiting student in InterACT lab, LTI, CMU, working with Prof. Stephan
Vogel.</p>
</blockquote>
<blockquote>
<p>I was awarded the Excellent Young Scholar Research Project by Jiangsu
Provincial Research Foundation in 2017, CCF-NLPCC Young Outstanding
Scientist Award in 2020, CIPSC Hanwang Youth Innovation Award in 2022.
Our alumni Hao Zhou, Zaixiang Zheng, Yu Bao (co-advised with
Prof. Jiajun Chen) were awarded Best Ph. D. Thesis Awards.</p>
</blockquote>
<blockquote>
<p>Our translation quality estimation research and systems won the first
place in WMT2022 QE subtasks (word-level MQM (En-De), sentence-level MQM
(En-De)) and all 3 tasks in En-De direction in WMT2023 QE (word-level
MQM, sentence-level MQM, error span detection).</p>
</blockquote>
<blockquote>
<p>Our paper is recognized as the Outstanding Paper in EMNLP 2024
(Congrats to Jiahuan and Yiqing!).</p>
</blockquote>
<hr />
<p><span id="recruiting"></p>
<h2 id="recruiting">Recruiting</h2>
<blockquote>
<p>I am looking for highly motivated undergraduate students to work
together on NLP problems. We are also expecting post-doc researchers to
work together on NLP research or applications. Please find <a
href="https://mp.weixin.qq.com/s/QHKMLzHW7cS0PoiIxjkcxQ">here</a> an
intro on Weixin.</p>
</blockquote>
<blockquote>
<p>For undergraduates, we give priority to an NLP background and an
summer project. Our summer project may invovle reading, discussion and
implementatin of our previous research, as well as further explorations.
Please get in touch before July.</p>
</blockquote>
<blockquote>
<p>If you have no NLP background, please consider taking our course (<a
href="mtcourse.html">Machine Translation and Natural Langauge
Generation</a>, every spring, for sophomores in Nanjing University) or
watching our <a href="https://www.bilibili.com/video/BV1TE421P7jp/">NLP
Summer Camp</a> (mainly for freshman or sophomore) videos first. There
are some talks about our research on bilibili.com (see <a
href="#presentations">Presentations</a>). Please also take a look before
you apply.</p>
</blockquote>
<blockquote>
<p>I am terribly sorry for not being able to reply all emails applying
for a regular Ph.D. position (overwhelmed by the applications). My reply
is usually fast. Please consider me as unavailable if no reply within
3~4 working days.</p>
</blockquote>
<hr />
<p><span id="interests"></p>
<h2 id="research-interests">Research Interests</h2>
<blockquote>
<p>My research is supported by projects from National Natural Science
Foundation of China (NSFC), National Key R&D Program of China and
the Jiangsu Provincial Research Foundation for Basic Research. We also
have wide collaborations with industrial labs in Baidu, Tencent,
Alibaba, ByteDance, Huawei, ZTE, China Mobile, etc.</p>
</blockquote>
<blockquote>
<p>My research interests lie in natural language processing (NLP), which
automatically understand and generate natural language texts, with
special focus on the language-related ability of LLMs (including
knowledge, reasoning, etc.) and the mutlingual aspect of LLMs (including
langauge transfer, alignment, etc.). We’re trying to exploring the
potential of LLMs by <strong>reinforcement learning</strong> (RL).
Compared with supervised training (such as instruction tuning), RL could
have a better balance between exploration and exploitation, which may
lead the LLMs to learn the internal mechanism of performing a task,
rather than some “short-cuts” to the answers.</p>
</blockquote>
<p>Currently, we mainly focus on the following topics:</p>
<blockquote>
<ul>
<li><strong>Language/Intelligent Abilities of LLM</strong>. We’re trying
to understand the abilities of LLMs, including the abilities for
language comprehension, language generation, in-context learning,
knowledge learning, logic reasoning, safety and alignment, as well as
the ability of working as intelligent agents for a given task (agentic
and long-trajectory reasoning). It is interesting and also important to
know how these abilities are obtained, how they are applied to specific
tasks and how to improve them if needed.</li>
</ul>
</blockquote>
<blockquote>
<ul>
<li><strong>LLM and Multilingualism (Multilingual LLM)</strong>. We’re
trying to understand and improve the relation of language abilities
across different languages in LLMs. One ideal way is to align different
languages altogether, in both knowledge and reasoning abilities. The
alginemnt may happen or be performed during the pretraining, continue
pretraining, or post training phase. In the end, we aim at building LLMs
that are equally effective for all languages.</li>
</ul>
</blockquote>
<blockquote>
<ul>
<li><strong>Machine Translation (MT)</strong>. We’re investigating
various methods to improve the capacity of multilingualism, by designing
novel MT architectures with the help of both task-specific models and
LLMs, evaluating/estimating the quality of MT, bringing human into the
learning loop, improving translation performance for smaller models,
specific scenarios, specific domains or low-resource language pairs,
etc.</li>
</ul>
</blockquote>
<blockquote>
<ul>
<li><strong>Interdisciplinary Application of NLP/NLG</strong>. We’re
actively applying the techniques of machine translation and natural
language processing in interdisciplinary research. For example, to
analyze ancient Chinese, to generate or translate ancient Chinese, to
understand the language comprehension process of human brain, to
modeling the structure of protein, etc.</li>
</ul>
</blockquote>
<hr />
<p><span id="services"></p>
<h2 id="academic-services">Academic Services</h2>
<h3 id="services-in-academic-society">Services in Academic Society</h3>
<ul>
<li><strong>Board Member</strong> of the Chinese Information Processing
Society of China (CIPSC) (2022-now)</li>
<li><strong>Deputy Director</strong> of the Technical Committee on
Machine Translation, of the Chinese Information Processing Society of
China (CIPSC) (2019-now)</li>
<li><strong>Executive Board Member</strong> of the Youth Working
Committee, of the Chinese Information Processing Society of China
(CIPSC) (2017-2020)</li>
</ul>
<h3 id="services-in-conferences-and-events">Services in Conferences and
Events</h3>
<ul>
<li><strong>Program Committee (co-)Chair</strong>: Program co-Chair for
CCMT2019, NLPCC2022; Demonstration co-Chair for CCL2020; Evaluation
Chair for CWMT (2017, 2018)</li>
<li><strong>Senior Program Committee Member</strong>: IJCAI (2020,
2021), AAAI (2022, 2024, 2026)</li>
<li><strong>Senior Area (co-)Chair</strong>: EACL2024, AACL2025,
ACL2026</li>
<li><strong>Senior Action Editor</strong>: ACL Rolling Review
(2023-)</li>
<li><strong>Area (co-)Chair</strong>: NLPCC2016, CCL2018, CCL2021,
ACL2021, EMNLP2022, EMNLP2023, EACL2023, ACL2023, CoLing2024, NAACL2024,
AAAI2027</li>
<li><strong>Program Committee Member/Reviewer</strong> for Major
conferences in NLP and AI: ACL, EMNLP, NAACL, AAAI, IJCAI, NeurIPS,
ICLR, ICML</li>
<li><strong>Committee Member</strong> for international machine
translation competitions: WMT2017, AI Challenger (2017,2018)</li>
<li><strong>International Journal Reviewer</strong>: TPAMI, TKDE, AI,
TACL, TALSP, TALLIP, WWWJ, FCS, JCST, NEUNET, IPM, KNOSYS, FMRE,
etc.</li>
</ul>
<hr />
<p><span id="honors"></p>
<h2 id="honors-and-awards">Honors and Awards</h2>
<ul>
<li><h4
id="excellent-young-scholar-research-project-by-jiangsu-provincial-research-foundation-2017">Excellent
Young Scholar Research Project by Jiangsu Provincial Research Foundation
(2017)</h4></li>
<li><h4
id="ccmt-best-english-paper-award-2019-ccmt-best-paper-award-2023">CCMT
Best English Paper Award (2019), CCMT Best Paper Award (2023)</h4></li>
<li><h4 id="cipsc-outstanding-services-award-2019">CIPSC Outstanding
Services Award (2019)</h4></li>
<li><h4 id="ccf-nlpcc-young-outstanding-scientist-award-2020">CCF-NLPCC
Young Outstanding Scientist Award (2020)</h4></li>
<li><h4 id="cipsc-hanwang-youth-innovation-award-2022">CIPSC Hanwang
Youth Innovation Award (2022)</h4></li>
<li><h4 id="jcst-outstanding-reviewer-award-2020-2022">JCST Outstanding
Reviewer Award (2020-2022)</h4></li>
</ul>
<ul>
<li><h4
id="jsai-science-and-technology-award-technology-innovation-award-2023">JSAI
Science and Technology Award (Technology Innovation Award,
2023)</h4></li>
<li><h4 id="emnlp-outstanding-paper-award-2024">EMNLP Outstanding Paper
Award (2024)</h4></li>
</ul>
<ul>
<li><h4 id="jscs-science-and-technology-award-first-price-2025">JSCS
Science and Technology Award (First Price, 2025)</h4></li>
<li><h4 id="icml-gold-reviewer-2026">ICML Gold Reviewer (2026)</h4></li>
</ul>
<h2 id="co-advisor-of-awarded-thesis">(Co-)advisor of Awarded
Thesis</h2>
<ul>
<li><h4
id="caai-best-ph.-d.-thesis-award-hao-zhou-2019-co-advised-with-prof.-jiajun-chen">CAAI
Best Ph. D. Thesis Award (Hao Zhou (2019, co-advised with Prof. Jiajun
Chen))</h4></li>
<li><h4
id="cipsc-best-ph.-d.-thesis-award-zaixiang-zheng-2022-co-advised-with-prof.-jiajun-chen">CIPSC
Best Ph. D. Thesis Award (Zaixiang Zheng (2022, co-advised with
Prof. Jiajun Chen))</h4></li>
<li><h4
id="excellent-master-thesis-in-jiangsu-province-rongxiang-weng-2020-co-advised-with-prof.-jiajun-chen-peng-wu-2020-co-advised-with-prof.-jianbing-zhang">Excellent
Master Thesis in Jiangsu Province (Rongxiang Weng (2020, co-advised with
Prof. Jiajun Chen), Peng Wu (2020, co-advised with Prof. Jianbing
Zhang))</h4></li>
<li><h4
id="jsai-best-ph.-d.-thesis-award-yu-bao-2022-co-advised-with-prof.-jiajun-chen">JSAI
Best Ph. D. Thesis Award (Yu Bao (2022, co-advised with Prof. Jiajun
Chen))</h4></li>
<li><h4
id="excellent-master-thesis-of-nanjing-university-or-the-departmentschool-level-rongxiang-weng-2020-peng-wu-2020-yanling-xiao-2022-dongqi-wang-2022-2023-yiming-yan-2023-zhijun-wang-2026">Excellent
Master Thesis of Nanjing University or the department/school level
(Rongxiang Weng (2020), Peng Wu (2020), Yanling Xiao (2022), Dongqi Wang
(2022, 2023), Yiming Yan (2023), Zhijun Wang (2026))</h4></li>
<li><h4
id="advisor-of-excellent-undergraduate-thesis-in-nanjing-university-renfei-dang-2025-second-prize">Advisor
of Excellent Undergraduate Thesis in Nanjing University (Renfei Dang
(2025, second prize))</h4></li>
</ul>
<hr />
<p><span id="courses"></p>
<h2 id="courses">Courses</h2>
<ul>
<li><h4 id="basic-programming.">Basic Programming.</h4>
<p>in Autumn semester, for undergraduate students.</p></li>
<li><h4 id="advanced-programming.">Advanced Programming.</h4>
<p>in Spring semester, for undergraduate students.</p></li>
<li><h4 id="programming-for-artificial-intelligence-intro.">Programming
for Artificial Intelligence (<a href="aipcourse.html">intro</a>).</h4>
<p>in Spring semester, for undergraduate students.</p></li>
<li><h4
id="natural-language-processing-machine-translation-part.">Natural
Language Processing (Machine Translation part).</h4>
<p>in Autumn semester, for graduate students.</p></li>
<li><h4
id="machine-translation-and-natural-language-generation-intro.">Machine
Translation and Natural Language Generation (<a
href="mtcourse.html">intro</a>).</h4>
<p>in Spring semester, for both undergraduate and graduate
students.</p></li>
<li><h4 id="can-machine-think-language-understanding-part-notes.">Can
Machine Think (Language Understanding part (<a
href="Language_Understanding/index.html">notes</a>)).</h4>
<p>in Spring semester, for undergraduate students.</p></li>
</ul>
<hr />
<h2 id="presentations">Presentations</h2>
<p><span id="presentations"></p>
<p>(mostly in Chinese, hosted on bilibili.com)</p>
<h4 id="research-talks-of-our-own-research">Research Talks (of our own
research)</h4>
<ul>
<li><p><strong>Unsupervised Learning of Machine Translation for
LLMs</strong> (TransZero)(30 mins, <a
href="https://www.bilibili.com/video/BV1qaKp63EVU/">video</a>). Lecture
for Summer School at NJU. 2026-07.</p></li>
<li><p><strong>Multilingual Ability of LLMs and Methods to Improve
It</strong> (CLiKA, SynCS, PreAlign, MoE-LPR, DeltaMoE)(65 mins, <a
href="https://www.bilibili.com/video/BV1toKm6wEnp/">video</a>). Lecture
for Summer School at NJU. 2026-07.</p></li>
<li><p><strong>Introduction to the Research of Multilingual
LLMs</strong> (70 mins, <a
href="https://www.bilibili.com/video/BV1zG1CBPEKZ/">video</a>). Invited
talk at BJTU. 2025-11.</p></li>
<li><p><strong>Knowledge Learning in Large Languae Models</strong> (80
mins, <a href="https://www.bilibili.com/video/BV1khSKYSEKT/">video</a>).
Invited talk at HIT-SZ. 2024-11.</p></li>
<li><p><strong>Multilingual LLM Research from an MT perspective</strong>
(20 mins, <a
href="https://www.bilibili.com/video/BV1zWzwYkEYD">video</a>). Invited
talk at CCMT-2024. 2024-11.</p></li>
<li><p><strong>Application of the Multilingual Ability of Large Language
Models</strong> (50 mins, <a
href="https://www.bilibili.com/video/BV1j7421Z77h/">video</a>). Invited
talk at HIT-SZ (later at Huawei, CLLM2024), 2024-05.</p></li>
<li><p><strong>Exploring the Multilingual Ability of Large Language
Models</strong> (60 mins, <a
href="https://www.bilibili.com/video/BV1Q94y1r7XN">video</a>). Invited
talk at Bytedance. 2023-08.</p></li>
<li><p><strong>Analyzing Multilingual Machine Translation Ability of
Large Language Models</strong> (30 min, <a
href="https://www.bilibili.com/video/BV16h411j7nW">video</a>). Invited
talk at HIT. 2023-04.</p></li>
<li><p><strong>Natural Language Generation with Latent
Variables</strong> (75 mins, <a
href="https://www.bilibili.com/video/BV1vu411v72L">video</a>). Invited
Talk at FDU. 2020-06.</p></li>
</ul>
<h4 id="tutorials-of-related-work">Tutorials (of related work)</h4>
<ul>
<li><p><strong>Research and Challenges of Multilingual Large Language
Models</strong> (50 mins, <a
href="https://www.bilibili.com/video/BV1G8UdYzE46">video</a>, <a
href="https://github.com/NJUNLP/tutorial_multilingual_LLM">slides</a>).
Tutorial Talk at NLPCC2024. 2024-10.</p></li>
<li><p><strong>Research Development of Multilingualism in the Era of
LLM</strong> (15 mins, <a
href="https://www.bilibili.com/video/BV125HPevEsn">video</a>). Invited
talk at the Frontier Research Overview session, CCL2024.
2024-07.</p></li>
<li><p><strong>Research Development of Machine translation and Large
Language Model</strong> (15 mins, <a
href="https://www.bilibili.com/video/BV1wu4y1b7rv">video</a>). Invited
talk at the Frontier Research Overview session, CCL2023.
2023-08.</p></li>
<li><p><strong>kNN Machine Translation</strong>. Tutorial at NLPCC2022
(75 mins, <a
href="https://www.bilibili.com/video/BV1XP411J7VA">video</a>, <a
href="https://github.com/NJUNLP/KNNMT-tutorial">slides</a>) and later at
MLNLP2022 (<a
href="https://www.bilibili.com/video/BV1b84y167Mb">video</a>).</p></li>
<li><p><strong>Quality Estimation for Machine Translation</strong> (100
mins, <a href="https://www.bilibili.com/video/BV1tz4y1a71R">video</a>).
Invited talk at MTPE workshop. 2022-02 (updated 2023-05).</p></li>
</ul>
<h4 id="introduction-for-new-comers">Introduction (for new comers)</h4>
<ul>
<li><p><strong>Large Language Models and Applications</strong> (2 hours,
<a href="https://www.bilibili.com/video/BV1TE421P7jp?p=6">video</a>).
Lecture at NJUNLP SummerCamp2024. 2024-07.</p></li>
<li><p><strong>Machine Translation and Multilingual LLMs</strong> (2
hours, <a
href="https://www.bilibili.com/video/BV1TE421P7jp?p=3">video</a>).
Lecture at NJUNLP SummerCamp2024. 2024-07.</p></li>
<li><p><strong>Introduction to Machine Translation</strong> (90 mins, <a
href="https://www.bilibili.com/video/BV1XX4y1n7Je?p=5">video</a>).
Lecture at NJUNLP SummerCamp 2023. 2023-07.</p></li>
</ul>
<hr />
<p><span id="pub"></p>
<h2 id="selected-publications">Selected Publications</h2>
<p><a
href="https://scholar.google.com.hk/citations?user=HF3-E9kAAAAJ&hl=en">A
More Complete List on Google Scholar</a></p>
<p>* marks corresponding author(s).</p>
<!-- --- FILTER BUTTONS --- -->
<div class="filter-tabs">
<p><strong>Filter by topic:</strong>
<button class="filter-btn active" data-filter="all">Show
All</button><br>
<button class="filter-btn" data-filter="MLLM">Multilingual LLM</button>,
including <button class="filter-btn" data-filter="MLLM-Align">Language
Alignment</button>
<button class="filter-btn" data-filter="MLLM-LowRes">Language
Enhancement</button>;<br>
<button class="filter-btn" data-filter="MT-">Machine
Translation</button>, including
<button class="filter-btn" data-filter="MT-LLM">LLM-MT</button>
<button class="filter-btn" data-filter="MT-QE">QE</button>
<button class="filter-btn" data-filter="MT-kNN">kNN</button>
<button class="filter-btn" data-filter="MT-IMT">Interactive MT</button>
<button class="filter-btn" data-filter="MT-NAT">NAT</button>;<br>
<button class="filter-btn" data-filter="LLM-">LLM Abilities</button>,
including
<button class="filter-btn" data-filter="LLM-Knowledge">Knowledge</button>
<button class="filter-btn" data-filter="LLM-Reasoning">Reasoning</button>
<button class="filter-btn" data-filter="LLM-Safety">Safety</button>;<br>
<button class="filter-btn" data-filter="X-">Interdiscipline</button>,
including
<button class="filter-btn" data-filter="X-Brain">Neuroscience</button>
<button class="filter-btn" data-filter="X-Bio">Protein</button>;<br>
<button class="filter-btn" data-filter="Generation">Generation</button>,
<button class="filter-btn" data-filter="NLP">other NLP
topics</button>.<br><br> <strong>Search by keywords:</strong>
<input type="text" id="search-input" placeholder="keywords..."></p>
</div>
<!-- --- END BUTTONS --- -->
<p>2026</p>
<div class="paper-item"
data-topic="LLM-Reasoning,MLLM-Align,MLLM-LowRes">
<blockquote>
<p><strong>[RP-OPSD]</strong> <strong>RP-OPSD: Reasoning-Pivot-Guided
On-Policy Self-Distillation for Multilingual Reasoning
Transfer</strong>. <br> Xinye Wang, Junxiao Liu, <strong>Shujian
Huang</strong>*. <br> preprint. (<a
href="https://arxiv.org/abs/2608.06347">arXiv:2608.06347</a>, <a
href="https://github.com/NJUNLP/RP-OPSD">code</a>)</p>
</blockquote>
</div>
<div class="paper-item"
data-topic="LLM-Reasoning,MLLM-Align,MLLM-LowRes">
<blockquote>
<p><strong>[PCS]</strong> <strong>Efficient Multilingual Reasoning
Transfer via Progressive Code-Switching</strong>. <br> Zhijun Wang,
Junxiao Liu, Hao Zhou, Hao-Ran Wei, Baosong Yang, <strong>Shujian
Huang</strong>*. <br> preprint. (<a
href="https://arxiv.org/abs/2607.00485">arXiv:2607.00485</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-QE,LLM-Reasoning,MT-LLM">
<blockquote>
<p><strong>[RIEQE]</strong> <strong>Unlocking Fine-Grained Translation
Quality Estimation in LRMs through Synergistically Evolving Implicit and
Explicit Reasoning</strong>. <br> Renfei Dang, Xinye Wang, Zhejian Lai,
Weilu Xu, Shimin Tao, Daimeng Wei, Min Zhang, <strong>Shujian
Huang</strong>*. <br> preprint. (<a
href="https://arxiv.org/abs/2605.31378">arXiv:2605.31378</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Safety">
<blockquote>
<p><strong>[ISA]</strong> <strong>Friend or Foe: How LLMs’ Safety Mind
Gets Fooled by Intent Shift Attack</strong>. <br> Peng Ding, Jun Kuang,
Wen Sun, Zongyu Wang, Xuezhi Cao, Xunliang Cai, Jiajun Chen,
<strong>Shujian Huang</strong>*. <br> Accepted by TACL. (<a
href="https://arxiv.org/abs/2511.00556">arXiv:2511.00556</a>, <a
href="https://github.com/NJUNLP/ISA">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="X-Bio">
<blockquote>
<p><strong>[DPLM-Evo]</strong> <strong>Towards A Generative Protein
Evolution Machine with DPLM-Evo</strong>.<br> Xinyou Wang, Liang Hong,
Jiasheng Ye, Zaixiang Zheng, Yu Li, <strong>Shujian Huang</strong>*,
Quanquan Gu*. <br> ICML 2026. (<a
href="https://arxiv.org/abs/2605.00182">arxiv:2605.00182</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning,MLLM-Align">
<blockquote>
<p><strong>[TAPO]</strong> <strong>TAPO: Translation Augmented Policy
Optimization for Multilingual Reasoning</strong>. <br> Xu Huang, Zhejian
Lai, Zixian Huang, Jiajun Chen, <strong>Shujian Huang</strong>*. <br>
preprint. (<a
href="https://arxiv.org/abs/2603.25419">arxiv:2603.25419</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning,MLLM">
<blockquote>
<p><strong>[ExpLang]</strong> <strong>ExpLang: Improved Exploration and
Exploitation in LLM Reasoning with On-Policy Thinking Language
Selection</strong>. <br> Changjiang Gao, Zixian Huang, Kaichen Yang,
Jiajun Chen, Jixing Li, <strong>Shujian Huang</strong>*. <br> preprint.
(<a href="https://arxiv.org/abs/2602.21887">arxiv:2602.21887</a>, <a
href="https://github.com/NJUNLP/ExpLang">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-LLM,LLM-Reasoning,MT-QE">
<blockquote>
<p><strong>[GRRM]</strong> <strong>GRRM: Group Relative Reward Modeling
for Machine Translation</strong>. <br> Sen Yang, Shanbo Cheng, Lu Xu,
Jianbing Zhang, <strong>Shujian Huang</strong>*. <br> preprint. (<a
href="https://arxiv.org/abs/2602.14028">arxiv:2602.14028</a>, <a
href="https://github.com/NJUNLP/GRRM">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning,MLLM-Align">
<blockquote>
<p><strong>[R3S]</strong> <strong>R3S: Refining and Recovering
Reinforcement Signals for Multilingual Understanding and
Reasoning</strong> ([TRIT] Self-Improving Multilingual Long Reasoning
via Translation-Reasoning Integrated Training). <br> Junxiao Liu, Zhijun
Wang, Yixiao Li, Zhejian Lai, Liqian Huang, Xin Huang*, Xue Han, Junlan
Feng, <strong>Shujian Huang</strong>*. <br> preprint. (<a
href="https://arxiv.org/abs/2602.05940">arxiv:2602.05940</a>, <a
href="https://github.com/NJUNLP/TRIT">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-LLM">
<blockquote>
<p><strong>Improving Long-Context Translation via Self-Supervised Dual
Learning</strong>. <br> Shanbo Cheng, Shuaijie She, Yu Bao, Jianbing
Zhang, Jiajun Chen, <strong>Shujian Huang</strong>*. <br> ACL 2026.</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-LowRes">
<blockquote>
<p><strong>[DeltaMoE]</strong> <strong>A Data-Efficient Path to
Multilingual LLMs: Language Expansion via Post-training PARAM
Integration into Upcycled MoE</strong>. <br> Hao Zhou, Tianhao Li,
Zhijun Wang, Shuaijie She, Linjuan Wu, Hao-Ran Wei*, Baosong Yang,
Jiajun Chen, <strong>Shujian Huang</strong>*. <br> ACL 2026. (<a
href="https://arxiv.org/abs/2605.18083">arxiv:2605.18083</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-LLM">
<blockquote>
<p><strong>[PEGRL]</strong> <strong>PEGRL: Improving Machine Translation
by Post-Editing Guided Reinforcement Learning</strong>. <br> Yunzhi
Shen, Hao Zhou, Xin Huang*, Xue Han, Junlan Feng, <strong>Shujian
Huang</strong>*. <br> Findings of ACL 2026. (<a
href="https://arxiv.org/abs/2602.03352">arxiv:2602.03352</a>, <a
href="https://github.com/NJUNLP/peg-rl">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning">
<blockquote>
<p><strong>Reasoning While Asking: Transforming Reasoning Large Language
Models from Passive Solvers to Proactive Inquirers</strong>. <br> Xin
Chen, Feng Jiang, Yiqian Zhang, Hardy Chen, Shuo Yan, Wenya Xie, Min
Yang*, <strong>Shujian Huang</strong>*. <br> ACL 2026. (<a
href="https://arxiv.org/abs/2601.22139">arxiv:2601.22139</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Knowledge">
<blockquote>
<p><strong>[KnownPatch]</strong> <strong>Understanding
New-Knowledge-Induced Factual Hallucinations in LLMs: Analysis and
Interpretation</strong>. <br> Renfei Dang, Peng Hu, Zhejian Lai,
Changjiang Gao, Min Zhang, <strong>Shujian Huang</strong>*. <br>
Findings of ACL 2026. (<a
href="https://arxiv.org/abs/2511.02626">arXiv:2511.02626</a>, <a
href="https://github.com/NJUNLP/New-Knowledge-Induced-Factual-Hallucinations">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning">
<blockquote>
<p><strong>[DuPO]</strong> <strong>DuPO: Enabling Reliable LLM
Self-Verification via Dual Preference Optimization</strong>. <br>
Shuaijie She, Yu Bao, Yu Lu, Lu Xu, Tao Li, Wenhao Zhu, <strong>Shujian
Huang</strong>*, Shanbo Cheng*, Lu Lu, Yuxuan Wang. <br> ICLR 2026. (<a
href="https://arxiv.org/abs/2508.14460">arxiv:2508.14460</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning">
<blockquote>
<p><strong>[InternalBias]</strong> <strong>The First Impression Problem:
Internal Bias Triggers Overthinking in Reasoning Models</strong>
(Internal Bias in Reasoning Models leads to Overthinking). <br> Renfei
Dang, Zhening Li, <strong>Shujian Huang</strong>*, Jiajun Chen. <br>
ICLR 2026. (<a
href="https://arxiv.org/abs/2505.16448">arxiv:2505.16448</a>, <a
href="https://github.com/NJUNLP/LongCoT-Internal-Bias">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>[LanguageRelatedNeurons]</strong> <strong>How does Alignment
Enhance LLMs’ Multilingual Capabilities? A Language Neurons
Perspective</strong>. <br> Shimao Zhang, Zhejian Lai, Xiang Liu,
Shuaijie She, Xiao Liu, Yeyun Gong, <strong>Shujian Huang</strong>*,
Jiajun Chen. <br> AAAI 2026. (<a
href="https://arxiv.org/abs/2505.21505">arXiv:2505.21505</a>, <a
href="https://github.com/NJUNLP/Language-Neurons-Alignment">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-LowRes">
<blockquote>
<p><strong>[TransLLM]</strong> <strong>Why Not Transform Chat Large
Language Models to Non-English?</strong>. <br> Xiang Geng, Ming Zhu,
Jiahuan Li, Zhejian Lai, Wei Zou, Shuaijie She, Jiaxin Guo, Xiaofeng
Zhao, Yinglu Li, Yuang Li, Chang Su, Yanqing Zhao, Xinglin Lyu, Min
Zhang, Jiajun Chen, Hao Yang, <strong>Shujian Huang</strong>*. <br>
Frontiers of Computer Science (2026, 20(7):2007356). (<a
href="https://arxiv.org/abs/2405.13923">arxiv:2405.13923</a>, <a
href="https://github.com/hy5468/TransLLM">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="Generation">
<blockquote>
<p><strong>[LDP]</strong> <strong>Improved Paraphrase Generation via
Controllable Latent Diffusion</strong>. <br> Wei Zou, Ziyuan Zhuang,
Xiang Geng, <strong>Shujian Huang</strong>*, Jia Liu, Jiajun Chen. <br>
Frontiers of Computer Science (2026, 20(1):2001302). (<a
href="https://arxiv.org/abs/2404.08938">arxiv:2404.08938</a>, <a
href="https://github.com/NJUNLP/ld4pg">code</a>)</p>
</blockquote>
</div>
<p>2025</p>
<div class="paper-item" data-topic="LLM-Reasoning">
<blockquote>
<p><strong>[AdaR]</strong> <strong>AdaR: A Framework for Equipping LLMs
with Adaptive Reasoning</strong> (Making Mathematical Reasoning
Adaptive). <br> Zhejian Lai, Xiang Geng, Zhijun Wang, Yang Bai, Jiahuan
Li, Rongxiang Weng, Jingang Wang, Xuezhi Cao, Xunliang Cai,
<strong>Shujian Huang</strong>*. <br> preprint. (<a
href="https://arxiv.org/abs/2510.04617">arXiv:2510.04617</a>, <a
href="https://github.com/NJUNLP/AdaR">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-LLM,MLLM">
<blockquote>
<p><strong>[EAxO]</strong> <strong>EnAnchored-X2X: English-Anchored
Optimization for Many-to-Many Translation</strong>. <br> Sen Yang, Yu
Bao, Yu Lu, Jiajun Chen, <strong>Shujian Huang</strong>*, Shanbo Cheng*.
<br> EMNLP 2025. (<a
href="https://arxiv.org/abs/2509.19770">arxiv:2509.19770</a>, <a
href="https://github.com/NJUNLP/EAX">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Safety">
<blockquote>
<p><strong>[SDGO]</strong> <strong>SDGO: Self-Discrimination-Guided
Optimization for Consistent Safety in Large Language Models</strong>.
<br> Peng Ding, Wen Sun, Dailin Li, Wei Zou, Jiaming Wang, Jiajun Chen,
<strong>Shujian Huang</strong>*. <br> EMNLP 2025. (<a
href="https://arxiv.org/abs/2508.15648">arxiv:2508.15648</a>, <a
href="https://github.com/NJUNLP/SDGO">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>[xMRC]</strong> <strong>Understanding LLMs’ Cross-Lingual
Context Retrieval: How Good It Is And Where It Comes From</strong>. <br>
Changjiang Gao, Hankun Lin, Xin Huang, Xue Han, Junlan Feng, Chao Deng,
Jiajun Chen, <strong>Shujian Huang</strong>*. <br> EMNLP 2025. (<a
href="https://arxiv.org/abs/2504.10906">arxiv:2504.10906</a>, <a
href="https://github.com/NJUNLP/Cross-Lingual-Context-Retrieval">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning">
<blockquote>
<p><strong>[R-PRM]</strong> <strong>R-PRM: Reasoning-Driven Process
Reward Modeling</strong>. <br> Shuaijie She, Junxiao Liu, Yifeng Liu,
Jiajun Chen, Xin Huang, <strong>Shujian Huang</strong>*. <br> EMNLP
2025. (<a href="https://arxiv.org/abs/2503.21295">arxiv:2503.21295</a>,
<a href="https://github.com/NJUNLP/R-PRM">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM,MT-LLM">
<blockquote>
<p><strong>[BenchMAX]</strong> <strong>BenchMAX: A Comprehensive
Multilingual Evaluation Suite for Large Language Models</strong>. <br>
Xu Huang, Wenhao Zhu, Hanxu Hu, Conghui He, Lei Li, <strong>Shujian
Huang</strong>*, Fei Yuan*. <br> Findings of EMNLP 2025. (<a
href="https://arxiv.org/abs/2502.07346">arxiv:2502.07346</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="X-Brain">
<blockquote>
<p><strong>Increasing alignment of large language models with language
processing in the human brain</strong>. <br> Changjiang Gao, Zhengwu Ma,
Jiajun Chen, Ping Li, <strong>Shujian Huang</strong>*, Jixing Li*. <br>
Nature Computational Science (IF:18.3). (<a
href="https://www.nature.com/articles/s43588-025-00863-0">nature
article</a>, <a
href="https://www.biorxiv.org/content/10.1101/2024.08.15.608196v4">biorxiv</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Safety">
<blockquote>
<p><strong>[SAGE]</strong> <strong>Why Not Act on What You Know?
Unleashing Safety Potential of LLMs via Self-Aware Guard
Enhancement</strong>. <br> Peng Ding, Jun Kuang, Zongyu Wang, Xuezhi
Cao, Xunliang Cai, Jiajun Chen, <strong>Shujian Huang</strong>*. <br>
Findings of ACL 2025. (<a
href="https://arxiv.org/abs/2505.12060">arxiv:2505.12060</a>, <a
href="https://github.com/NJUNLP/SAGE">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="X-Bio">
<blockquote>
<p><strong>[DPLM-2.1]</strong> <strong>Elucidating the Design Space of
Multimodal Protein Language Models</strong>. <br> Cheng-Yen Hsieh,
Xinyou Wang, Daiheng Zhang, Dongyu Xue, Fei Ye, <strong>Shujian
Huang</strong>, Zaixiang Zheng, Quanquan Gu*. <br> ICML 2025.
<strong>spotlight</strong> (<a
href="https://arxiv.org/abs/2504.11454">arxiv:2504.11454</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM,MT-LLM">
<blockquote>
<p><strong>[Trans-Zero]</strong> <strong>TRANS-ZERO: Self-Play
Incentivizes Large Language Models for Multilingual Translation Without
Parallel Data</strong>. <br> Wei Zou, Sen Yang, Yu Bao, <strong>Shujian
Huang</strong>*, Jiajun Chen, Shanbo Cheng*. <br> Findings of ACL 2025.
(<a href="https://arxiv.org/abs/2504.14669">arxiv:2504.14669</a>, <a
href="https://github.com/NJUNLP/trans0">code</a>, <a
href="https://github.com/NJUNLP/trans0-NPU">code-npu-version</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM">
<blockquote>
<p><strong>Could Thinking Multilingually Empower LLM
Reasoning?</strong>. <br> Changjiang Gao, Xu Huang, Wenhao Zhu,
<strong>Shujian Huang</strong>*, Lei Li, Fei Yuan*. <br> preprint. (<a
href="https://arxiv.org/abs/2504.11833">arxiv:2504.11833</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>[SynCS]</strong> <strong>Investigating and Scaling up
Code-Switching for Multilingual Language Model Pre-Training</strong>.
<br> Zhijun Wang, Jiahuan Li, Hao Zhou, Rongxiang Weng, Jingang Wang,
Xin Huang, Xue Han, Junlan Feng, Chao Deng, <strong>Shujian
Huang</strong>*. <br> Findings of ACL 2025. (<a
href="https://arxiv.org/abs/2504.01801">arxiv:2504.01801</a>, <a
href="https://github.com/NJUNLP/SynCS">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning">
<blockquote>
<p><strong>Process-based Self-Rewarding Language Models</strong>. <br>
Shimao Zhang, Xiao Liu*, Xin Zhang, Junxiao Liu, Zheheng Luo,
<strong>Shujian Huang</strong>*, Yeyun Gong. <br> Findings of ACL 2025.
(<a href="https://arxiv.org/abs/2503.03746">arxiv:2503.03746</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-QE">
<blockquote>
<p><strong>[DCSQE]</strong> <strong>Alleviating Distribution Shift in
Synthetic Data for Machine Translation Quality Estimation</strong>. <br>
Xiang Geng, Zhejian Lai, Jiajun Chen, Hao Yang, <strong>Shujian
Huang</strong>*. <br> ACL 2025. (<a
href="https://arxiv.org/abs/2502.19941">arXiv:2502.19941</a>, <a
href="https://github.com/NJUNLP/njuqe">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM">
<blockquote>
<p><strong>Generalizing From Short to Long: Effective Data Synthesis for
Long-Context Instruction Tuning</strong>. <br> Wenhao Zhu, Pinzhen Chen,
Hanxu Hu, <strong>Shujian Huang</strong>*, Fei Yuan, Jiajun Chen,
Alexandra Birch. <br> preprint. (<a
href="https://arxiv.org/abs/2502.15592">arxiv:2502.15592</a>, <a
href="https://github.com/NJUNLP/context-synthesis">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-LLM">
<blockquote>
<p><strong>[SelfEvo]</strong> <strong>Self-Evolution Knowledge
Distillation for LLM-based Machine Translation</strong>. <br> Yuncheng
Song, Liang Ding, Changtong Zan, <strong>Shujian Huang</strong>*. <br>
COLING 2025. (<a
href="https://arxiv.org/abs/2412.15303">arxiv:2412.15303</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Knowledge,LLM-Reasoning">
<blockquote>
<p><strong>[CKFR]</strong> <strong>Large Language Models Are
Cross-Lingual Knowledge-Free Reasoners</strong>. <br> Peng Hu, Sizhe
Liu, Changjiang Gao, Xin Huang, Xue Han, Junlan Feng, Chao Deng,
<strong>Shujian Huang</strong>*. <br> NAACL 2025. (<a
href="https://arxiv.org/abs/2406.16655">arxiv:2406.16655</a>, <a
href="https://www.bilibili.com/video/BV1khSKYSEKT/">video</a>, <a
href="https://github.com/NJUNLP/Knowledge-Free-Reasoning">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="X-Bio">
<blockquote>
<p><strong>[DPLM-2]</strong> <strong>DPLM-2: A Multimodal Diffusion
Protein Language Model</strong>. <br> Xinyou Wang, Zaixiang Zheng, Fei
YE, Dongyu Xue, <strong>Shujian Huang</strong>, Quanquan Gu*. <br> ICLR
2025. (<a
href="https://arxiv.org/abs/2410.13782">arxiv:2410.13782</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-LowRes">
<blockquote>
<p><strong>[MoE-LPR]</strong> <strong>MoE-LPR: Multilingual Extension of
Large Language Models through Mixture-of-Experts with Language Priors
Routing</strong>. <br> Hao Zhou, Zhijun Wang, <strong>Shujian
Huang</strong>*, Xin Huang, Xue Han, Junlan Feng, Chao Deng, Weihua Luo,
Jiajun Chen. <br> AAAI 2025. (<a
href="https://arxiv.org/abs/2408.11396">arxiv:2408.11396</a>, <a
href="https://github.com/NJUNLP/MoE-LPR">code</a>, <a
href="https://github.com/NJUNLP/MoE-LPR-NPU">code-npu-version</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM,Generation">
<blockquote>
<p><strong>Multi-Candidate Speculative Decoding</strong>. <br> Sen Yang,
<strong>Shujian Huang</strong>*, Xinyu Dai, Jiajun Chen. <br> NLPCC
2025. (<a href="https://arxiv.org/abs/2401.06706">arxiv:2401.06706</a>,
<a href="https://github.com/NJUNLP/MCSD">code</a>)</p>
</blockquote>
</div>
<p>2024</p>
<div class="paper-item" data-topic="LLM-Knowledge">
<blockquote>
<p><strong>Formality is Favored: Unraveling the Learning Preferences of
Large Language Models on Data with Conflicting Knowledge</strong>. <br>
Jiahuan Li, Yiqing Cao, <strong>Shujian Huang</strong>*, Jiajun Chen.
<br> EMNLP 2024. <strong>Outstanding Paper Award</strong>
(5+20/1271/6105). (<a
href="https://arxiv.org/abs/2410.04784">arxiv:2410.04784</a>, <a
href="https://www.bilibili.com/video/BV1khSKYSEKT/">video</a>, <a
href="https://github.com/NJUNLP/Formality-is-Favored">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Safety">
<blockquote>
<p><strong>[Hallu-PI]</strong> <strong>Hallu-PI: Evaluating
Hallucination in Multi-modal Large Language Models within Perturbed
Inputs</strong>. <br> Peng Ding, Jingyu Wu, Jun Kuang, Dan Ma, Xuezhi
Cao, Xunliang Cai, Shi Chen, Jiajun Chen, <strong>Shujian
Huang</strong>*. <br> ACM MM 2024. (<a
href="https://arxiv.org/abs/2408.01355">arxiv:2408.01355</a>, <a
href="https://github.com/NJUNLP/Hallu-PI">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>[PreAlign]</strong> <strong>PreAlign: Boosting Cross-Lingual
Transfer by Early Establishment of Multilingual Alignment</strong>. <br>
Jiahuan Li, <strong>Shujian Huang</strong>*, Xinyu Dai, Jiajun Chen.
<br> EMNLP 2024. (<a
href="https://arxiv.org/abs/2407.16222">arxiv:2407.16222</a>, <a
href="https://www.bilibili.com/video/BV1khSKYSEKT/">video</a>, <a
href="https://github.com/NJUNLP/PreAlign">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-LowRes">
<blockquote>
<p><strong>Multilingual Contrastive Decoding via Language-Agnostic
Layers Skipping</strong>.<br> Wenhao Zhu, Sizhe Liu, <strong>Shujian
Huang</strong>*, Shuaijie She, Chris Wendler, Jiajun Chen. <br> Findings
of EMNLP 2024. (<a
href="https://arxiv.org/abs/2407.10795">arxiv:2407.10795</a>, <a
href="https://github.com/NJUNLP/SkipLayerCD">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Reasoning,LLM-Knowledge">
<blockquote>
<p><strong>[OCKR]</strong> <strong>Large Language Models are Limited in
Out-of-Context Knowledge Reasoning</strong>. (OCKR)<br> Peng Hu,
Changjiang Gao, Ruiqi Gao, Jiajun Chen, <strong>Shujian Huang</strong>*.
<br> Findings of EMNLP 2024. (<a
href="https://arxiv.org/abs/2406.07393">arxiv:2406.07393</a>, <a
href="https://github.com/NJUNLP/ID-OCKR">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>Getting More from Less: Large Language Models are Good
Spontaneous Multilingual Learners</strong>. <br> Shimao Zhang,
Changjiang Gao, Wenhao Zhu, Jiajun Chen, Xin Huang, Xue Han, Junlan
Feng, Chao Deng, <strong>Shujian Huang</strong>*. <br> EMNLP 2024. (<a
href="https://arxiv.org/abs/2405.13816">arxiv:2405.13816</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>The Power of Question Translation Training in Multilingual
Reasoning: Broadened Scope and Deepened Insights</strong>. <br> Wenhao
Zhu, <strong>Shujian Huang</strong>*, Fei Yuan, Jiajun Chen, Alexandra
Birch. <br> preprint. (<a
href="https://arxiv.org/abs/2405.01345">arxiv:2405.01345</a>, <a
href="https://github.com/NJUNLP/QAlign">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="X-Bio">
<blockquote>
<p><strong>[DPLM]</strong> <strong>Diffusion Language Models Are
Versatile Protein Learners</strong>. <br> Xinyou Wang, Zaixiang Zheng,
Fei Ye, Dongyu Xue, <strong>Shujian Huang</strong>, Quanquan Gu*. <br>
ICML 2024. (<a
href="https://arxiv.org/abs/2402.18567">arxiv:2402.18567</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="X-Brain">
<blockquote>
<p><strong>Measuring Meaning Composition in the Human Brain with
Composition Scores from Large Language Models</strong>. <br> Changjiang
Gao, Jixing Li*, Jiajun Chen, <strong>Shujian Huang</strong>*. <br> ACL
2024. (<a href="https://arxiv.org/abs/2403.04325">arxiv:2403.04325</a>,
<a
href="https://github.com/RiverGao/ffn_composition_analysis">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align,LLM-Reasoning">
<blockquote>
<p><strong>[QAlign]</strong> <strong>Question Translation Training for
Better Multilingual Reasoning</strong>. <br> Wenhao Zhu, <strong>Shujian
Huang</strong>*, Fei Yuan, Shuaijie She, Jiajun Chen, Alexandra Birch.
<br> Findings of ACL 2024. (<a
href="https://arxiv.org/abs/2401.07817">arxiv:2401.07817</a>, <a
href="https://github.com/NJUNLP/QAlign">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align,LLM-Reasoning">
<blockquote>
<p><strong>[MAPO]</strong> <strong>MAPO: Advancing Multilingual
Reasoning through Multilingual Alignment-as-Preference
Optimization</strong>. <br> Shuaijie She, Wei Zou, <strong>Shujian
Huang</strong>*, Wenhao Zhu, Xiang Liu, Xiang Geng, Jiajun Chen. <br>
ACL 2024. (<a
href="https://arxiv.org/abs/2401.06838">arxiv:2401.06838</a>, <a
href="https://github.com/NJUNLP/MAPO">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-QE">
<blockquote>
<p><strong>Lost in the Source Language: How Large Language Models
Evaluate the Quality of Machine Translation</strong>. <br> Xu Huang,
Zhirui Zhang*, Xiang Geng, Yichao Du, Jiajun Chen, <strong>Shujian
Huang</strong>*. <br> Findings of ACL 2024. (<a
href="https://arxiv.org/abs/2401.06568">arxiv:2401.06568</a>, <a
href="https://github.com/NJUNLP/lost_in_the_src">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM-Align">
<blockquote>
<p><strong>[CLiKA]</strong> <strong>Multilingual Pretraining and
Instruction Tuning Improve Cross-Lingual Knowledge Alignment, But Only
Shallowly</strong>. <br> Changjiang Gao, Hongda Hu, Peng Hu, Jiajun
Chen, Jixing Li, <strong>Shujian Huang</strong>*.<br> NAACL 2024. (<a
href="https://arxiv.org/abs/2403.04325">arxiv:2403.04325</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MT-LLM">
<blockquote>
<p><strong>[MT-PATCHER]</strong> <strong>MT-PATCHER: Selective and
Extendable Knowledge Distillation from Large Language Models for Machine
Translation</strong>. <br> Jiahuan Li, Shanbo Cheng, <strong>Shujian
Huang</strong>*, Jiajun Chen.<br> NAACL 2024. (<a
href="https://arxiv.org/abs/2403.09522">arxiv:2403.09522</a>, <a
href="https://github.com/NJUNLP/MT-Patcher">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Safety">
<blockquote>
<p><strong>[ReNeLLM]</strong> <strong>A Wolf in Sheep’s Clothing:
Generalized Nested Jailbreak Prompts can Fool Large Language Models
Easily</strong>. <br> Peng Ding, Jun Kuang, Dan Ma, Xuezhi Cao, Yunsen
Xian, Jiajun Chen, <strong>Shujian Huang</strong>*.<br> NAACL 2024. (<a
href="https://arxiv.org/abs/2311.08268">arxiv:2311.08268</a>, <a
href="https://github.com/NJUNLP/ReNeLLM">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="LLM-Knowledge,LLM-Reasoning">
<blockquote>
<p><strong>Exploring the Factual Consistency in Dialogue Comprehension
of Large Language Models</strong>. <br> Shuaijie She, <strong>Shujian
Huang</strong>*, Xingyun Wang, Yanke Zhou, Jiajun Chen. <br> NAACL 2024.
(<a href="https://arxiv.org/abs/2311.07194">arxiv:2311.07194</a>, <a
href="https://github.com/NJUNLP/CoP">code</a>)</p>
</blockquote>
</div>
<div class="paper-item" data-topic="MLLM,MT-LLM">
<blockquote>
<p><strong>Multilingual Machine Translation with Large Language Models:
Empirical Results and Analysis</strong>. <br> Wenhao Zhu, Hongyi Liu,
Qingxiu Dong, Jingjing Xu, <strong>Shujian Huang</strong>*, Lingpeng
Kong, Jiajun Chen, Lei Li. <br> Findings of NAACL 2024. (<a
href="https://arxiv.org/abs/2304.04675">arxiv:2304.04675</a>, <a
href="https://www.bilibili.com/video/BV16h411j7nW">video</a>, <a