-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
887 lines (800 loc) · 48 KB
/
index.html
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
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta name=viewport content="width=800">
<meta name="generator" content="HTML Tidy for Linux/x86 (vers 11 February 2007), see www.w3.org">
<style type="text/css">
/* Color scheme stolen from Sergey Karayev */
a {
color: #1772d0;
text-decoration: none;
}
a:focus,
a:hover {
color: #f09228;
text-decoration: none;
}
body,
td,
th,
tr,
p,
a {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 15px
}
strong {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 15px;
}
heading {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 22px;
}
papertitle {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 15px;
font-weight: 700
}
name {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 32px;
}
.one {
width: 160px;
height: 120px;
position: relative;
}
span.highlight {
background-color: #ffffd0;
}
</style>
<link rel="icon" type="image/png" href="images/logo12.png">
<title>Revanth Gangi Reddy</title>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<link href='https://fonts.googleapis.com/css?family=Lato:400,700,400italic,700italic' rel='stylesheet' type='text/css'>
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-167314770-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-167314770-1');
</script>
</head>
<body>
<table width="800" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="67%" valign="middle">
<p align="center">
<name>Revanth Gangi Reddy</name>
</p>
<p>I am a PhD student at the University of Illinois, Urbana Champaign, advised by <a href="http://blender.cs.illinois.edu/hengji.html" target="_blank">Prof. Heng Ji</a>. Previously, I was an AI Resident at IBM Research, New York wherein I had the pleasure of working with <a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-vittorio" target="_blank">Vittorio Castelli</a>, <a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avirup Sil</a> and <a href="https://www.ibm.com/ibm/ideasfromibm/us/ibm_fellows/2016/salim_roukos.html" target="_blank">Salim Roukos</a>.
</p>
<p> I graduated from Indian Institute of Technology Madras in 2018 with a bachelors degree in computer science. While at IIT Madras, I worked with <a href="https://www.cse.iitm.ac.in/~miteshk/" target="_blank">Prof. Mitesh Khapra</a> and <a href="https://www.cse.iitm.ac.in/~ravi/" target="_blank">Prof. Balaraman Ravindran</a> </p>
<p align=center>
<a href="mailto:g.revanthreddy111@gmail.com" target="_blank">Email</a>  / 
<a href="data/Resume_Revanth.pdf" target="_blank">CV</a>  / 
<a href="https://www.linkedin.com/in/revanth-gangi-reddy-5b7257ba/" target="_blank">LinkedIn</a>  / 
<a href="https://scholar.google.com/citations?user=SXP5Ej0AAAAJ&hl=en" target="_blank">Google Scholar</a>
</p>
</td>
<td width="33%">
<img style="width:100%;max-width:100%" src="images/RevanthGangiReddy.jpg">
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="100%" valign="middle">
<heading>Research</heading>
<p>
I'm interested in natural language processing and large language models, particulary in the fields of agentic search, embedding models and reranking, and retrieval-augmented generation.
</p>
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="100%" valign="middle">
<heading>Papers</heading>
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="25%">
<div class="one">
<a href="images/infogent.png" target="_blank"><img style="width:100%;max-width:100%" src='images/infogent.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://gangiswag.github.io/infogent/" target="_blank">
<papertitle>INFOGENT: An Agent-Based Framework for Web Information Aggregation</papertitle>
</a>
<br>
<strong>Revanth Reddy*</strong>,
<a href="https://sagnikmukherjee.github.io/" target="_blank">Sagnik Mukherjee*</a>,
<a href="https://wjdghks950.github.io/" target="_blank">Jeonghwan Kim*</a>,
<a href="https://mikewangwzhl.github.io/" target="_blank">Zhenhailong Wang*</a>,
<a href="https://siebelschool.illinois.edu/about/people/all-faculty/dilek" target="_blank">Dilek Hakkani-Tur</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>Under Review</em>
<br>
[<a href="https://www.arxiv.org/pdf/2410.19054" target="_blank">Paper</a>][<a href="https://gangiswag.github.io/infogent/" target="_blank">Blog Post</a>][<a href="https://github.com/gangiswag/infogent" target="_blank">Code</a>]
<br>
<p>We introduce INFOGENT,a novel modular and feedback-driven framework for web information aggregation involving three distinct components: Navigator, Extractor and Aggregator.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/cornstack.png" target="_blank"><img style="width:100%;max-width:100%" src='images/cornstack.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://gangiswag.github.io/code-embed/" target="_blank">
<papertitle>CoRNStack: High-Quality Contrastive Data for Better Code Retrieval and Reranking</papertitle>
</a>
<br>
<a href="https://tarsur909.github.io/" target="_blank">Tarun Suresh*</a>,
<strong>Revanth Reddy*</strong>,
<a href="" target="_blank">Yifei Xu</a>,
<a href="" target="_blank">Zach Nussbaum</a>,
<a href="" target="_blank">Andriy Mulyar</a>,
<a href="" target="_blank">Brandon Duderstadt</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>Under Review</em>
<br>
[<a href="https://gangiswag.github.io/cornstack" target="_blank">Blog Post</a>][<span>Paper Coming Soon!</span>][<a href="https://github.com/gangiswag/cornstack" target="_blank">Code</a>]
<br>
<p>We introduce CoRNStack, a large-scale, high-quality contrastive training dataset for code that spans multiple programming languages. We demonstrate that contrastive training of embedding models using CoRNStack leads to state-of-the-art performance across a variety of code retrieval tasks.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/SmartBook.png" target="_blank"><img style="width:100%;max-width:100%" src='images/SmartBook.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2303.14337" target="_blank">
<papertitle>SmartBook: AI-Assisted Situation Report Generation</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Daniel Lee</a>,
<a href="https://yrf1.github.io" target="_blank">Yi R. Fung</a>,
<a href="https://vickizeng.com" target="_blank">Qi Zeng</a>,
<a href="https://limanling.github.io" target="_blank">Manling Li</a>,
<a href="https://wzq016.github.io" target="_blank">Ziqi Wang</a>,
<a href="" target="_blank">Paul Sullivan</a>,
<a href="https://ieeexplore.ieee.org/author/37357721400" target="_blank">Clare Voss</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>Under Review</em>
<br>
[<a href="https://arxiv.org/pdf/2303.14337.pdf" target="_blank">Paper</a>][<a href="https://github.com/blender-nlp/SmartBook" target="_blank">Code</a>]
<br>
<p>We introduce SmartBook, a generalizable automated framework designed to assist human analysts in real-time situation report generation from large news corpora, by generating a structured report with multiple hypotheses (claims) summarized and grounded with rich links to factual evidence.</p>
</td>
</tr>
<tr>
<td width="25%" style="text-align: center;">
<div class="one">
<a href="images/FIRST.png" target="_blank"><img style="width:57%;max-width:100%" src='images/FIRST.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2406.15657" target="_blank">
<papertitle>FIRST: Faster Improved Listwise Reranking with Single Token Decoding</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">JaeHyeok Doo</a>,
<a href="" target="_blank">Yifei Xu</a>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank"> Md Arafat Sultan</a>,
<a href="" target="_blank">Deevya Swain</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avirup Sil</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>EMNLP 2024</em>
<br>
[<a href="https://arxiv.org/pdf/2406.15657" target="_blank">Paper</a>][<a href="https://github.com/gangiswag/llm-reranker" target="_blank">Code</a>]
<br>
<p>We introduce FIRST, a novel listwise LLM reranking approach leveraging the output logits of the first generated identifier to directly obtain a ranked ordering of the candidates.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/AGRAME.png" target="_blank"><img style="width:100%;max-width:100%" src='images/AGRAME.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2405.15028" target="_blank">
<papertitle>AGRaME: Any-Granularity Ranking with Multi-Vector Embeddings</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Omar Attia</a>,
<a href="https://yunyaoli.github.io" target="_blank">Yunyao Li</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>,
<a href="https://salonipotdar.github.io" target="_blank">Saloni Potdar</a>
<br>
<em>EMNLP 2024</em>
<br>
[<a href="https://arxiv.org/pdf/2405.15028" target="_blank">Paper</a>]
<br>
<p>We introduce any-granularity ranking which leverages multi-vector embeddings to rank at varying levels of granularity while maintaining encoding at a single (coarser) level of granularity.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/OpenQA.png" target="_blank"><img style="width:100%;max-width:100%" src='images/OpenQA.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://aclanthology.org/2024.findings-naacl.48/" target="_blank">
<papertitle>Towards Better Generalization in Open-Domain Question Answering by Mitigating Context Memorization</papertitle>
</a>
<br>
<a href="https://zhangzx-uiuc.github.io" target="_blank">Zixuan Zhang</a>,
<strong>Revanth Reddy</strong>,
<a href="http://www.kevinsmall.org" target="_blank">Kevin Small</a>,
<a href="https://tongzhang-ml.org" target="_blank">Tong Zhang</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>NAACL Findings, 2024</em>
<br>
[<a href="https://aclanthology.org/2024.findings-naacl.48.pdf" target="_blank">Paper</a>]
<br>
<p>We propose to improve an OpenQA model's generalizability across different corpora and domains by mitigating the model's over-memorization of knowledge.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/ReFIT.png" target="_blank"><img style="width:100%;max-width:100%" src='images/ReFIT.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2305.11744" target="_blank">
<papertitle>Inference-time Reranker Relevance Feedback for Neural Information Retrieval</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://pdasigi.github.io" target="_blank">Pradeep Dasigi</a>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank"> Md Arafat Sultan</a>,
<a href="https://armancohan.com" target="_blank">Arman Cohan</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<a href="https://homes.cs.washington.edu/~hannaneh/" target="_blank">Hannaneh Hajishirzi</a>
<br>
<em>Under Review</em>
<br>
[<a href="https://arxiv.org/pdf/2305.11744" target="_blank">Paper</a>]
<br>
<p>We propose to compute an improved vector representation of the query using supervision from the re-ranker at inference time, thereby improving the retriever's Recall@K. Our approach is parameter-free, lightweight, and can serve arbitrary retrieve-and-rerank pipelines, significantly improving retrieval recall in multiple domains, languages, and modalities.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/PersonaDB.png" target="_blank"><img style="width:100%;max-width:100%" src='images/PersonaDB.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2402.11060" target="_blank">
<papertitle>Persona-DB: Efficient Large Language Model Personalization for Response Prediction with Collaborative Data Refinement</papertitle>
</a>
<br>
<a href="https://chenkaisun.github.io" target="_blank">Chenkai Sun</a>,
<a href="https://empathyang.github.io" target="_blank">Ke Yang</a>,
<strong>Revanth Reddy</strong>,
<a href="https://yrf1.github.io" target="_blank">Yi R. Fung</a>,
<a href="https://www.fst.um.edu.mo/personal/hpchan/" target="_blank"> Hou Pong Chan</a>,
<a href="http://www.kevinsmall.org" target="_blank">Kevin Small</a>,
<a href="https://czhai.cs.illinois.edu" target="_blank">ChengXiang Zhai</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>Under Review</em>
<br>
[<a href="https://arxiv.org/pdf/2402.11060" target="_blank">Paper</a>]
<br>
<p>We introduce a framework that enhances the accuracy and context efficiency of retrieval-based LLM personalization through collaborative data refinement. The method also excels in cold-start scenarios.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/SocialBot.png" target="_blank"><img style="width:100%;max-width:100%" src='images/SocialBot.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="data/WSDM_2024_Demo_Alexa_SocialBot.pdf" target="_blank">
<papertitle>Progressive Responses with Real-Time Internet Search for Knowledge-Powered Conversations</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Sharath Chandra</a>,
<a href="" target="_blank">Hao Bai</a>,
<a href="" target="_blank"> Wentao Yao</a>,
<a href="" target="_blank"> Mankeerat Singh Sidhu</a>,
<a href="" target="_blank"> Karan Aggarwal</a>,
<a href="" target="_blank"> Prathamesh Sonawane</a>,
<a href="https://czhai.cs.illinois.edu" target="_blank">ChengXiang Zhai</a>
<br>
<em>WSDM Demo, 2024</em>
<br>
[<a href="data/WSDM_2024_Demo_Alexa_SocialBot.pdf" target="_blank">Paper</a>]
<br>
<p>We introduce the use of progressive response generation to integrate real-time web search results, where the preliminary
response buys time for a detailed follow-up, ensuring a smooth user interaction. As a result, our method cuts down user
waiting times for voice-based chatbots by 50%.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/EMNLP_Dialog_Query.png" target="_blank"><img style="width:100%;max-width:100%" src='images/EMNLP_Dialog_Query.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2310.14340v1" target="_blank">
<papertitle>Social Commonsense-Guided Search Query Generation for Open-Domain Knowledge-Powered Conversations</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Hao Bai</a>,
<a href="" target="_blank"> Wentao Yao</a>,
<a href="" target="_blank">Sharath Chandra</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<a href="https://czhai.cs.illinois.edu" target="_blank">ChengXiang Zhai</a>
<br>
<em>Findings of EMNLP, 2023</em>
<br>
[<a href="https://arxiv.org/pdf/2310.14340.pdf" target="_blank">Paper</a>]
<br>
<p>To tackle passive conversations, we propose to integrate social commonsense reasoning for the generation of search queries in knowledge-powered conversations. We leverage a commonsense dialog system to establish connections related to the conversation topic, which subsequently guides an instruction-driven query generation model. </p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/SumREN.png" target="_blank"><img style="width:100%;max-width:100%" src='images/SumREN.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/pdf/2212.01146.pdf" target="_blank">
<papertitle>SumREN: Summarizing Reported Speech about Events in News</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://www.amazon.science/author/heba-elfardy" target="_blank">Heba Elfardy</a>,
<a href="https://www.fst.um.edu.mo/personal/hpchan/" target="_blank"> Hou Pong Chan</a>,
<a href="http://www.kevinsmall.org" target="_blank">Kevin Small</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>AAAI, 2023</em>
<br>
[<a href="https://arxiv.org/pdf/2212.01146.pdf" target="_blank">Paper</a>][<a href="data/AAAI_poster___SumREN.pdf" target="_blank">Poster</a>]
<br>
<p>We propose the novel task of summarizing the reactions of different speakers with respect to a given event. We create a new multi-document summarization benchmark, SumREN, along with a pipeline-based framework for summarizing reported speech, which generates summaries that are more abstractive and factually consistent.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/NewsClaims.png" target="_blank"><img style="width:100%;max-width:100%" src='images/NewsClaims.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2112.08544" target="_blank">
<papertitle>NewsClaims: A New Benchmark for Claim Detection from News with Attribute Knowledge</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Sai Chinthakindi</a>,
<a href="https://mikewangwzhl.github.io" target="_blank"> Zhenhailong Wang</a>,
<a href="https://yrf1.github.io" target="_blank">Yi R. Fung</a>,
<a href="" target="_blank">Kathryn S. Conger</a>,
<a href="" target="_blank">Ahmed S. Elsayed</a>,
<a href="https://www.colorado.edu/faculty/palmer-martha/" target="_blank">Martha Palmer</a>,
<a href="https://mbzuai.ac.ae/study/faculty/preslav-nakov" target="_blank">Preslav Nakov</a>,
<a href="http://www.cs.cmu.edu/~hovy" target="_blank">Eduard Hovy</a>,
<a href="http://www.kevinsmall.org" target="_blank">Kevin Small</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>EMNLP, 2022</em>
<br>
[<a href="https://arxiv.org/pdf/2112.08544.pdf" target="_blank">Paper</a>][<a href="data/EMNLP_poster_NewsClaims.pdf" target="_blank">Poster</a>]
<br>
<p>We present NewsClaims, a new benchmark for knowledge-aware claim detection, that re-defines the claim detection problem to include extraction of additional attributes related to the claim. NewsClaims aims to benchmark claim detection in emerging scenarios, comprising unseen topics with no training data.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/IR_attn.png" target="_blank"><img style="width:100%;max-width:100%" src='images/IR_attn.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2204.11373" target="_blank">
<papertitle>Entity-Conditioned Question Generation for Robust Attention Distribution in Neural Information Retrieval</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank">Arafat Sultan</a>,
<a href="https://www.linkedin.com/in/martin-franz-138a7810/" target="_blank">Martin Franz</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>SIGIR, 2022</em>
<br>
[<a href="https://arxiv.org/pdf/2204.11373.pdf" target="_blank">Paper</a>][<a href="data/SIGIR_poster_2022.pdf" target="_blank">Poster</a>]
<br>
<p>Using a novel targeted synthetic data generation method that identifies poorly attended entities and conditions the generation episodes on those, we teach neural IR to attend more uniformly and robustly to all entities in a given passage.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/DPR_syn.png" target="_blank"><img style="width:100%;max-width:100%" src='images/DPR_syn.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2104.07800" target="_blank">
<papertitle>Towards Robust Neural Retrieval Models with Synthetic Pre-Training</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://sites.google.com/email.arizona.edu/vikasy95/home" target="_blank">Vikas Yadav</a>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank">Arafat Sultan</a>,
<a href="https://www.linkedin.com/in/martin-franz-138a7810/" target="_blank">Martin Franz</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-vittorio" target="_blank">Vittorio Castelli</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>
<br>
<em>COLING, 2022</em>
<br>
[<a href="https://arxiv.org/pdf/2104.07800.pdf" target="_blank">Paper</a>]
<br>
<p>We show that synthetic examples generated using a sequence-to-sequence generator can be effective in improving the robustness of neural IR, with gains in both in-domain and out-of-domain scenarios.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/QA_Claim_Framework.png" target="_blank"><img style="width:100%;max-width:100%" src='images/QA_Claim_Framework.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://aclanthology.org/2022.coling-1.603/" target="_blank">
<papertitle>A Zero-Shot Claim Detection Framework using Question Answering</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Sai Chinthakindi</a>,
<a href="https://yrf1.github.io" target="_blank">Yi R. Fung</a>,
<a href="http://www.kevinsmall.org" target="_blank">Kevin Small</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>COLING, 2022</em>
<br>
[<a href="https://aclanthology.org/2022.coling-1.603.pdf" target="_blank">Paper</a>][<a href="data/COLING_poster_ClaimQA.pdf" target="_blank">Poster</a>]
<br>
<p> We propose a fine-grained claim detection framework that leverages zero-shot question answering using directed questions to solve a diverse set of sub-tasks such as topic filtering, claim object detection, and claimer detection.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/MuMuQA.png" target="_blank"><img style="width:100%;max-width:100%" src='images/MuMuQA.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2112.10728">
<papertitle>MuMuQA: Multimedia Multi-Hop News Question Answering via Cross-Media Knowledge Extraction and Grounding</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="" target="_blank">Xilin Rui</a>,
<a href="https://limanling.github.io" target="_blank">Manling Li</a>,
<a href="https://xudonglinthu.github.io" target="_blank">Xudong Lin</a>,
<a href="https://www.haoyangwen.com" target="_blank">Haoyang Wen</a>,
<a href="https://j-min.io" target="_blank">Jaemin Cho</a>,
<a href="https://wilburone.github.io" target="_blank">Lifu Huang </a>,
<a href="https://www.cs.unc.edu/~mbansal/" target="_blank">Mohit Bansal</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>,
<a href="https://www.ee.columbia.edu/~sfchang/" target="_blank">Shih-Fu Chang</a>,
<a href="https://www.alexander-schwing.de" target="_blank">Alexander Schwing</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>AAAI, 2022</em>
<br>
[<a href="https://arxiv.org/pdf/2112.10728" target="_blank">Paper</a>][<a href="data/MuMuQA_AAAI.pdf" target="_blank">Slides</a>][<a href="data/AAAI_poster_MuMuQA.pdf" target="_blank">Poster</a>]
<br>
<p> We propose a new benchmark for multimedia question answering over news articles and introduce a novel data generation framework for generating questions that are grounded on objects in images and answered using the news body text.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/COVID_demo.png" target="_blank"><img style="width:100%;max-width:100%" src='images/COVID_demo.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://aclanthology.org/2022.acl-demo.13/" target="_blank">
<papertitle>COVID-19 Claim Radar: A Structured Claim Extraction and Tracking System</papertitle>
</a>
<br>
<a href="https://limanling.github.io" target="_blank">Manling Li</a>,
<strong>Revanth Reddy</strong>,
<a href="https://wzq016.github.io" target="_blank">Ziqi Wang</a>,
<a href="https://yschiangg.gitlab.io" target="_blank">Yi-Shyuan Chiang</a>,
<a href="https://laituan.io" target="_blank">Tuan M. Lai</a>,
<a href="" target="_blank">Pengfei Yu</a>,
<a href="https://zhangzx-uiuc.github.io" target="_blank">Zixuan Zhang</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>
<br>
<em>ACL Demo, 2022 </em>
<br>
[<a href="https://aclanthology.org/2022.acl-demo.13.pdf" target="_blank">Paper</a>][<a href="http://18.221.187.153/" target="_blank">Demo</a>]
<br>
<p> We present COVID-19 Claim Radar, a system that automatically extracts claims relating to COVID-19 in news articles. We provide a comprehensive structured view of such claims, with rich attributes (such as claimers and their affiliations) and associated knowledge elements (such as events, relations and entities).</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/COVID_QA.png" target="_blank"><img style="width:100%;max-width:100%" src='images/COVID_QA.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2204.09248" target="_blank">
<papertitle>Synthetic Target Domain Supervision for Open Retrieval QA</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-bsiyer" target="_blank">Bhavani Iyer</a>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank">Arafat Sultan</a>,
<a href="" target="_blank">Rong Zhang</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-vittorio" target="_blank">Vittorio Castelli</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-raduf" target="_blank">Radu Florian</a>,
<a href="https://www.ibm.com/ibm/ideasfromibm/us/ibm_fellows/2016/salim_roukos.html" target="_blank">Salim Roukos</a>
<br>
<em>SIGIR, 2021</em>
<br>
[<a href="https://arxiv.org/pdf/2204.09248.pdf" target="_blank">Paper</a>][<a href="data/SIGIR_poster.pdf" target="_blank">Poster</a>][<a href="data/CovidQA_SIGIR.pdf" target="_blank">Slides</a>]
<br>
<p>We explore using a synthetic example generation approach to improve the performance of state-of-the-art open-domain end-to-end question answering systems in a specialized domain, such as COVID-19.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/InfoSurgeon.png" target="_blank"><img style="width:100%;max-width:100%" src='images/InfoSurgeon.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://blender.cs.illinois.edu/paper/infosurgeon2021.pdf" target="_blank">
<papertitle>InfoSurgeon: Cross-Media Fine-grained Information Consistency Checking for Fake News Detection</papertitle>
</a>
<br>
<a href="http://yrf1.github.io" target="_blank">Yi R. Fung</a>,
<a href="http://people.cs.pitt.edu/~chris/" target="_blank">Chris Thomas</a>,
<strong>Revanth Reddy</strong>,
<a href="http://sandeep06011991.github.io" target="_blank">Sandeep Polisetty</a>,
<a href="https://blender.cs.illinois.edu/hengji.html" target="_blank">Heng Ji</a>,
<a href="https://www.ee.columbia.edu/~sfchang/" target="_blank">Shih-Fu Chang</a>,
<a href="http://www.cs.columbia.edu/~kathy/" target="_blank">Kathleen McKeown</a>,
<a href="https://www.cs.unc.edu/~mbansal/" target="_blank">Mohit Bansal</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>
<br>
<em>ACL, 2021</em>
<br>
[<a href="https://blender.cs.illinois.edu/paper/infosurgeon2021.pdf" target="_blank">Paper</a>]
<br>
<p>While most previous work is on document level fake news detection, for the first time we propose misinformation detection at knowledge element level. It not only achieves higher detection accuracy but also makes the results more explainable.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/KBQA.png" target="_blank"><img style="width:100%;max-width:100%" src='images/KBQA.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2012.01707" target="_blank">
<papertitle>Leveraging Abstract Meaning Representation for Knowledge Base Question Answering</papertitle>
</a>
<br>
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-kapanipa" target="_blank">Pavan Kapanipathi*</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=ibm-Ibrahim.Abdelaziz1" target="_blank">Ibrahim Abdelaziz*</a>,
<a href="https://www.linkedin.com/in/srinivasravishankar/" target="_blank">Srinivas Ravishankar*</a>,
... ,
<strong>Revanth Reddy</strong>,
<a href="https://scholar.google.com/citations?user=CqqheH0AAAAJ&hl=en" target="_blank">Ryan Riegel</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=ibm-Gaetano.Rossiello" target="_blank">Gaetano Rossiello</a>,
<a href="https://www.linkedin.com/in/uditsharma7/?originalSubdomain=in" target="_blank">Udit Sharma</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=in-gpshri27" target="_blank">Shrivatsa Bhargav</a>,
<a href="https://sites.google.com/site/moyunlp/" target="_blank">Mo Yu</a>
<br>
<em>Findings of ACL, 2021</em>
<br>
[<a href="https://arxiv.org/pdf/2012.01707.pdf" target="_blank">Paper</a>]
<br>
<p>We introduce a neuro-symbolic question answering system that leverages AMR for question understanding and uses a pipeline-based approach involving a semantic parser, entity and relationship linkers and a neuro-symbolic reasoner.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/TechQA.png" target="_blank"><img style="width:100%;max-width:100%" src='images/TechQA.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2010.05904" target="_blank">
<papertitle>Multi-Stage Pretraining for Low-Resource Domain Adaptation</papertitle>
</a>
<br>
<a href="" target="_blank">Rong Zhang*</a>,
<strong>Revanth Reddy*</strong>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank">Arafat Sultan</a>,
<a href="https://www.linkedin.com/in/efsuns/" target="_blank">Efsun Kayi</a>,
<a href="https://www.linkedin.com/in/anthony-ferritto-4bb65912a" target="_blank">Anthony Ferrito</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-vittorio" target="_blank">Vittorio Castelli</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>,
<a href="" target="_blank">Todd Ward</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-raduf" target="_blank">Radu Florian</a>,
<a href="https://www.ibm.com/ibm/ideasfromibm/us/ibm_fellows/2016/salim_roukos.html" target="_blank">Salim Roukos</a>
<br>
<em>EMNLP, 2020</em>
<br>
[<a href="https://arxiv.org/pdf/2010.05904.pdf" target="_blank">Paper</a>]
<br>
<p>We formulate synthetic pre-training tasks that can transfer to downstream tasks, by using structure in unlabeled text. We show considerable gains on multiple tasks in the IT domain: question answering, document ranking and duplicate question detection.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/Span-Correction.png" target="_blank"><img style="width:100%;max-width:100%" src='images/Span-Correction.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2011.03435" target="_blank">
<papertitle>Answer Span Correction in Machine Reading Comprehension</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://www.linkedin.com/in/mdarafatsultan" target="_blank">Arafat Sultan</a>,
<a href="" target="_blank">Rong Zhang</a>,
<a href="https://www.linkedin.com/in/efsuns/" target="_blank">Efsun Kayi</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-vittorio" target="_blank">Vittorio Castelli</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-avi" target="_blank">Avi Sil</a>
<br>
<em>Findings of EMNLP, 2020</em>
<br>
[<a href="https://arxiv.org/pdf/2011.03435.pdf" target="_blank">Paper</a>]
<br>
<p>We propose an approach for correcting partial match answers (EM=0, 0<F1<1) into exact match (EM=1, F1=1) and obtain upto 1.3% improvement over a RoBERTa-based machine reading comprehension system in both monolingual and multilingual evaluation.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/Mining.png" target="_blank"><img style="width:100%;max-width:100%" src='images/Mining.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/abs/2010.10673" target="_blank">
<papertitle>Pushing the Limits of AMR Parsing with Self-Learning</papertitle>
</a>
<br>
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-ysuklee" target="_blank">Young-suk Lee*</a>,
<a href="https://ramon-astudillo.github.io" target="_blank">Ramon Astudillo*</a>,
<a href="https://www.linkedin.com/in/tahira-naseem-12066b46" target="_blank">Tahira Naseem*</a>,
<strong>Revanth Reddy*</strong>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=us-raduf" target="_blank">Radu Florian</a>,
<a href="https://www.ibm.com/ibm/ideasfromibm/us/ibm_fellows/2016/salim_roukos.html" target="_blank">Salim Roukos</a>
<br>
<em>Findings of EMNLP, 2020</em>
<br>
[<a href="https://arxiv.org/pdf/2010.10673.pdf" target="_blank">Paper</a>]
<br>
<p>We propose self-learning approaches to improve AMR parsers, via generation of synthetic text and synthetic AMR as well as refinement of actions from the oracle. We achieve state-of-the-art performance in AMR parsing on benchmark AMR 1.0 and AMR 2.0 datasets.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/multi-level-memory.png" target="_blank"><img style="width:100%;max-width:100%" src='images/multi-level-memory.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://www.aclweb.org/anthology/N19-1375.pdf" target="_blank">
<papertitle>Multi-Level Memory for Task Oriented Dialogs</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://sites.google.com/site/danishcontractor1/home" target="_blank">Danish Contractor</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=in-diraghu1" target="_blank">Dinesh Raghu</a>,
<a href="https://researcher.watson.ibm.com/researcher/view.php?person=in-jsachind" target="_blank">Sachindra Joshi</a>
<br>
<em>NAACL, 2019</em>
<br>
[<a href="https://www.aclweb.org/anthology/N19-1375.pdf" target="_blank">Paper</a>][<a href="data/NAACL_poster.pdf" target="_blank">Poster</a>]
<br>
<p>We design a novel multi-level memory architecture that retains natural hierarchy of the knowledge base without breaking it down into subject-relation-object triples. We use separate memories for dialog context and KB to learn different memory readers.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/FigureNet.png" target="_blank"><img style="width:100%;max-width:100%" src='images/FigureNet.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://arxiv.org/pdf/1806.04655.pdf" target="_blank">
<papertitle>FigureNet : A Deep Learning model for Question-Answering on Scientific Plots</papertitle>
</a>
<br>
<strong>Revanth Reddy</strong>,
<a href="https://www.linkedin.com/in/rahul13ramesh" target="_blank">Rahul Ramesh</a>,
<a href="https://ameet-1997.github.io" target="_blank">Ameet Deshpande</a>,
<a href="https://www.cse.iitm.ac.in/~miteshk/" target="_blank">Mitesh Khapra</a>
<br>
<em>IJCNN, 2019</em>
<br>
[<a href="https://arxiv.org/pdf/1806.04655.pdf" target="_blank">Paper</a>][<a href="data/FigureNet.pdf" target="_blank">Slides</a>]
<br>
<p> We design a modular network that uses depth-wise and 1D convolutions for visual reasoning on scientific plots. We achieve state-of-the-art accuracy on FigureQA dataset, bettering Relation Networks by 7%, with a training time over an order of magnitude lesser.</p>
</td>
</tr>
<tr>
<td width="25%">
<div class="one">
<a href="images/ERG.png" target="_blank"><img style="width:100%;max-width:100%" src='images/ERG.png'></a>
</div>
</td>
<td valign="top" width="75%">
<a href="https://epubs.siam.org/doi/pdf/10.1137/1.9781611975673.40" target="_blank">
<papertitle>Edge Replacement Grammars : A Formal Language Approach for Generating Graphs</papertitle>
</a>
<br>
<strong>Revanth Reddy*</strong>,
<a href="http://www.sarathchandar.in" target="_blank">Sarath Chandar*</a>,
<a href="https://www.cse.iitm.ac.in/~ravi/" target="_blank">Balaraman Ravindran</a>
<br>
<em>SDM, 2019</em>
<br>
[<a href="https://arxiv.org/pdf/1902.07159.pdf" target="_blank">Paper</a>][<a href="data/SDM_Presentation.pdf" target="_blank">Slides</a>][<a href="data/SDM_poster.pdf" target="_blank">Poster</a>]
<br>
<p>We propose a graph generative model based on probabilistic edge replacement grammars. We design an algorithm to build graph grammars by capturing the statistically significant sub-graph patterns.</p>
</td>
</tr>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td>
<br>
<p align="right">
<font size="2">
<br><a href="https://jonbarron.info/">Template from here</a></font>
</p>
</td>
</tr>
</table>
</td>
</tr>
</table>
</body>
</html>