Ehsan Shareghi
Associate Professor in Natural Language Processing
Works on predictive models for language in text and speech, including reasoning in large language models.
Ehsan Shareghi's lab at UCL works on large language models and reasoning & agents. 39 papers since 2023, 13 at top venues, mostly EMNLP and Findings of EMNLP. 5 PhD students. 24 alumni.
At a glance
- papers, 2023–2026
- 10/ year
- at T1 venues
- 33%
- PhD students
- 5
- MS students
- 0
- alumni listed
- 24
Papers by yearfrom the lab homepage, 2023–2026
- 10 papers2026 · T1 4 · collecting
- 10 papers2025 · T1 5
- 13 papers2024 · T1 3
- 6 papers2023 · T1 1
T1 venuesOtherVenues: EMNLP, Findings of EMNLP, Findings of ACL, ACL, CoNLL
Research focuspaper titles that use each topic's words
- Large language models29 papers
- language16 papers
- Reasoning & agents10 papers
- text1 paper
- predictive0 papers
Paper list
From the lab homepage.
202610 papers
- Findings of EMNLPBeyond Surface Forms: Symbolic Edits as a Test for Logical Reasoning with LLMs
- CoNLLCould language models win the International Linguistics Olympiad?
- EMNLP (T1)Cube Bench: A Controlled Diagnostic for Spatial and Sequential Reasoning in Multimodal Large Language Models
- EMNLP (T1)GTS: Inference-Time Scaling of Latent Reasoning with a Learnable Gaussian Thought Sampler
- AAAI WorkshopImproving Symbolic Translation of Language Models for Logical Reasoning
- Artificial Intelligence and LawLegal Citation Prediction with LLMs: A Comparative Evaluation of Instruction Tuning, Retrieval, and Jurisdiction-Specific Pre-training on the AusLaw Citation Benchmark
- ACL (T1)Privacy-R1: Privacy-Aware Multi-LLM Agent Collaboration via Reinforcement Learning
- Findings of ACLTowards Inference-time Scaling for Continuous Space Reasoning
- COLMTRIDENT: Benchmarking LLM Safety in Finance, Medicine, and Law
- AAAI (T1)Uncertainty-Based Methods for Automated Process Reward Data Construction and Output Aggregation in Mathematical Reasoning
202510 papers
- Findings of EMNLPAssessing the Sensitivity and Alignment of FOL Closeness Metrics
- NAACL (T1)Audio Is the Achilles' Heel: Red Teaming Audio Large Multimodal Models
- Findings of EMNLPDiscrete Minds in a Continuous World: Do Language Models Know Time Passes?
- EMNLP (T1)Graph-Based Confidence Estimation for Large Language Model Reasoning
- COLMJigsaw Puzzles: Splitting Harmful Questions to Jailbreak Large Language Models
- EMNLP (T1)Logical Reasoning with Outcome Reward Models for Test-Time Scaling
- ICML (T1)Measuring, Evaluating and Improving Logical Consistency in Large Language Models
- EMNLP (T1)Reshaping Representation Space to Balance Safety and Over-rejection in Large Audio Language Models
- International Speech Communication AssociationSpeechDialogueFactory: Generating High-Quality Speech Dialogue Data to Accelerate Your Speech-LLM Development
- Findings of EMNLPVerifiAgent: a Unified Verification Agent in Language Model Reasoning
202413 papers
- ALTAA Closer Look at Logical Reasoning with LLMs: The Choice of Tool Matters
- COLMAligning with Human Judgement: The Role of Pairwise Preference in Large Language Model Evaluators
- Findings of EMNLPCan LLMs Reason in the Wild with Programs?
- ALTAConversational SIMULMT: Efficient Simultaneous Translation with Large Language Models
- EACLEquipping Language Models with Tool Use Capability for Tabular Data Analysis in Finance
- Findings of EMNLPExploring the Potential of Multimodal LLM with Knowledge-Intensive Multimodal ASR
- ACL (T1)Harnessing the Power of Large Language Models for Natural Language to First-Order Logic Translation
- Oxford Intersections: AI in SocietyNot Explainable but Verifiable: Alternative First Steps in Overcoming the Problems Associated with AI’s Answers to Legal Problems
- Findings of ACLPiVe: Prompting with Iterative Verification to Improve Graph-based Generative Capability of LLMs
- Findings of EACLReward Engineering for Generating Semi-structured Explanation
- EMNLP (T1)Towards Probing Speech-Specific Risks in Large Multimodal Models: A Taxonomy, Benchmark, and Insights
- Findings of ACLTowards Uncertainty-Aware Language Agent
20236 papers
- CoNLLA Minimal Approach for Natural Language Action Space in Text-based Games
- Findings of EACLGenerating Synthetic Speech from SpokenVocab for Speech Translation
- International Speech Communication AssociationInvestigating Pre-trained Audio Encoders in the Low-Resource Condition
- EMNLP (T1)Koala: An Index for Quantifying Overlaps with Pre-training Corpora
- CogSciOn Reality and the Limits of Language Data: Aligning LLMs with Human Norms
- Findings of EMNLPPOSQA: Probe the World Models of LLMs with Size Comparisons
Where alumni went24 listed
Apply to UCLFall 2027
Not announced yet
Computer Science MPhil/PhD
- Round
- Main
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNTsinghua
Yang Liu
- Multilingual, translation & linguistics
- GBCambridge
Fermin Moscoso del Prado Martin
- Multilingual, translation & linguistics
- GBUCL
Mehrnoosh Sadrzadeh
- Multilingual, translation & linguistics
- USMichigan
Steve Abney
- Multilingual, translation & linguistics
- CNPeking
Dongyan Zhao
- Knowledge, QA & retrieval
- DETUM
Alexander Fraser
- Multilingual, translation & linguistics
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.