Ashutosh Adhikari
@aadhikariii
Followers
223
Following
1K
Media
8
Statuses
116
PhD Student at @EdinburghNLP Previously at Microsoft Turing; CS @UWaterloo; @Mila_Quebec Views my own.
Edinburgh, Scotland
Joined June 2015
Excited to share my first work as a PhD student at @EdinburghNLP that I will be presenting at EMNLP! RQ1: Can we achieve scalable oversight across modalities via debate? Yes! We show that debating VLMs lead to better model quality of answers for reasoning tasks.
1
7
13
Got burned by an Apple ICLR paper โ it was withdrawn after my Public Comment. So hereโs what happened. Earlier this month, a colleague shared an Apple paper on arXiv with me โ it was also under review for ICLR 2026. The benchmark they proposed was perfectly aligned with a
55
218
2K
@iatitov and I are happy to announce a fully-funded 3.5-year PhD studentship at @EdinburghNLP, for September 2026 start, on language-based state representations for time series modelling, with applications in health data monitoring and beyond. {1/5}
1
15
31
Hey thatโs @EdinburghNLP!
CSRankings counts publication in top conferences to rank professors/universities. But this encourages researchers to pursue quantity rather than quality. We propose https://t.co/ajtH10DbWQ, a new university ranking system that tries to measure quality instead of quantity of
4
2
51
At #EMNLP2025 in Suzhou to present the last work of my PhDโจ ๐
tomorrow ๐ 12.30 session We studied temporal reasoning on legal events and show that LLMs struggle with the linguistic complexities of legal language. Joint work between @EdinburghNLP and @TechAtBloomberg
0
1
4
Can multimodal LLMs truly understand research poster images?๐ ๐ We introduce PosterSumโa new multimodal benchmark for scientific poster summarization! ๐ชง ๐ Dataset: https://t.co/B5NzvqnWUA ๐ Paper: https://t.co/EHt4SwaGF3
3
27
82
I will be at EMNLP next week presenting this work on November the 7th! Reach out to me for any questions :)) Work done with my advisor, Mirella Lapata! Preprint: https://t.co/eu6tPw7YyP
#EMNLP2025 #multimodallearning
0
0
0
RQ3: Where do debate or consultancy fail? Our analysis show that judges benefit when the experts are arguing for diverse opinions! Red quadrant is when the judge is persuaded more often than they should (i.e. they are deceptive).
1
0
0
RQ2: Can debate be used as a reliable mechanism for yielding quality reasoning data? Yes! We show that the reasoning data attained from debate in a completely unsupervised manner imbue reasoning in the expert vision language models.
1
0
1
Considering a PhD/MSc in NLP? Iโm hiring students this cycle! If you are passionate about making language models reliable and safe, eager about understanding and controlling language models, and would like to add to your research some multilingual flavor - apply to my group! ๐
16
102
737
Hello everyone. A friend told me that I shouldn't post this message because it made me and other PhD students look bad. But I actually think it's important to show how PhD students (especially foreign ones) have to deal not only with research-related difficulties, but also with
gofundme.com
Hello everyone. A friend told me that I shouldn't post this messageโฆ Mathieu Alain needs your support for Help me resuming my PhD studies after sick leave
44
245
1K
Hi folks, I am embarrassed to say my account was compromised for a couple of weeks and X took forever to verify me. I want to apologize for any inappropriate content you might have come across from my account during the period.
0
0
1
Seems fitting to repost ๐
8
13
299
By popular demand we've extended the Wordplay Workshop deadline by a couple of weeks until Sept 12! The competition on realistic dialogue for game agents already has over 5000 submissions and the winners will also be at the workshop. Come hang out with us at EMNLP!
The Wordplay Workshop is back! 5th edition with EMNLP in Suzhou this Dec. We're also hosting a competition this time on making more realistic LLM powered NPCs in games! As always come by and chat all things text agents!
2
8
16
๐๏ธ Lifetime Achievement Award at #ACL2025NLP A standing ovation for Prof. Kathy McKeown, recipient of the ACL 2025 Lifetime Achievement Award! ๐
1
17
125
Instead of complaining that peer review is dead, take a positive step to improve it today. The reviewers are not aliens, they are us! - Revise your review and make it clear. Identify the crucial points that impacted your score negatively and positively. - If the paper is
10
14
153
Looking for 2 emergency ACs for ACL ARR May cycle (EMNLP 2025) to oversee 2-3 papers in Conversational AI that touch on reasoning/multimodality. You should have published in EMNLP or similar venues and ideally have 5 years of experience reviewing. Email me if interested. thanks!
4
4
24
Preprint: Can we learn to reason for story generation (~100k tokens), without reward models? Yes! We introduce an RLVR-inspired reward paradigm VR-CLI that correlates with human judgements of quality on the 'novel' task of Next-Chapter Prediction. Paper: https://t.co/eO0nUHzRjG
7
51
326
My fitness age after commencing my PhD : 18-> 21 in a matter of 4 months. Help!
0
0
0
(1/5) How should you represent characters for long-story tasks? Our EMNLP 2024 paper proposes a new character-sheet format, called CHIRON, that can be used for both downstream tasks and story analysis. arxiv: https://t.co/YGPKYuMjAF code: https://t.co/oWP7Ur9mFf
2
6
22