Papers by Shay Cohen
Can Large Language Model Summarizers Adapt to Diverse Scientific Communication Goals? (2024.findings-acl)
Copied to clipboard
| Challenge: | Recent work on the evaluation of large language models (LLMs) has shown unprecedented performance on diverse language generation tasks. |
| Approach: | They investigate the controllability of large language models on scientific summarization tasks by controlling stylistic and content coverage factors. |
| Outcome: | The proposed model outperforms humans on the MuP review generation task in terms of similarity to reference summaries and human preferences. |
Evaluating Automatic Metrics with Incremental Machine Translation Systems (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies have shown that neural metrics are more reliable than non-neural metrics. |
| Approach: | They propose to use commercial machine translations to evaluate machine translation metrics based on their preference for more recent outputs. |
| Outcome: | The proposed dataset confirms several previous findings, including the advantage of neural metrics over non-neural ones, and also explores the debated issue of how MT quality affects metric reliability. |
Modeling News Interactions and Influence for Financial Market Prediction (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies often adopt a simplified approach by treating available news data holistically and investigating its overall effect on market outcomes, the nuanced information contained within individual news items is overlooked. |
| Approach: | They propose a market prediction model that integrates multi-modal information from both market data and news articles to capture the links between news and prices. |
| Outcome: | The proposed model outperforms existing market prediction models with an improvement of 0.429 and 0.341 in the daily Sharpe ratio for the two markets respectively. |
Co-training an Unsupervised Constituency Parser with Weak Supervision (2022.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for unsupervised parsing that use bootstrapping classifiers to identify if a node dominates a span are lacking. |
| Approach: | They propose a method for unsupervised parsing that relies on bootstrapping classifiers to identify if a node dominates a specific span. |
| Outcome: | The proposed method achieves 63.1 F1 on the English test set and new state-of-the-art on treebanks for Chinese and Japanese. |
LeanReasoner: Boosting Complex Logical Reasoning with Lean (2024.naacl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) often struggle with complex logical reasoning due to logical inconsistencies and the inherent difficulty of such reasoning. |
| Approach: | They propose a method that formalizes logical reasoning problems into theorems within Lean and then proves or disproving the corresponding theorels. |
| Outcome: | The proposed method achieves state-of-the-art performance on the FOLIO dataset and near this level on ProofWriter. |
A Joint Matrix Factorization Analysis of Multilingual Representations (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies have demonstrated that pre-trained models acquire and incorporate linguistic knowledge in their multilingual representations. |
| Approach: | They propose a tool for comparing latent representations of multilingual and monolingual models . they use joint matrix factorization to analyze multiple sets of representations in a joint manner . |
| Outcome: | The proposed tool analyzes latent representations of multilingual and monolingual models . it shows that language properties influence the factorization outputs . |
AMR Parsing is Far from Solved: GrAPES, the Granular AMR Parsing Evaluation Suite (2023.emnlp-main)
Copied to clipboard
| Challenge: | Abstract Meaning Representation parsers have improved in recent years, but not solved. |
| Approach: | They propose an evaluation suite that evaluates AMR parsers on a range of phenomena . they find that current parser outputs are far from satisfactory . |
| Outcome: | The proposed evaluation suite reveals the abilities and shortcomings of current parsers. |
Detecting and Mitigating Hallucinations in Multilingual Summarisation (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing faithfulness metrics for abstractive summarisation models focus on English . metric mFACT is best suited to detect hallucinations in cross-lingual transfer . |
| Approach: | They propose a method to evaluate the faithfulness of non-English summaries by translation-based transfer from multiple English faithfulness metrics. |
| Outcome: | The proposed method reduces hallucinations in cross-lingual transfer by weighing the loss of each training example by its faithfulness score. |
Layer by Layer: Uncovering Where Multi-Task Learning Happens in Instruction-Tuned Large Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Pre-trained large language models retain task-specific knowledge, but where and to what extent they retain it remains unexplored. |
| Approach: | They investigate the task-specific information encoded in pre-trained LLMs and the effects of instruction tuning on their representations across over 60 NLP tasks. |
| Outcome: | The results show that pre-trained models retain task-specific knowledge . some tasks are already encoded in pre-train models, but others benefit from instruction tuning. |
Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions (2024.emnlp-main)
Copied to clipboard
| Challenge: | Using a method to identify next-token neurons, we find that some attention heads recognize contexts relevant to predicting a token and activate a downstream token-predicting neuron accordingly. |
| Approach: | They propose a method to identify next-token neurons and determine the upstream attention heads responsible for their activity in LLMs. |
| Outcome: | The proposed method identifies next-token neurons, finds prompts that highly activate them, and determines the upstream attention heads responsible. |
Large Language Models Relearn Removed Concepts (2024.findings-acl)
Copied to clipboard
| Challenge: | a recent study has shown that models can re-acquire pruned concepts after editing . however, it remains unclear whether models can recover such concepts after retraining . |
| Approach: | They evaluate concept relearning in large language models by tracking concept saliency and similarity in pruned neurons during retraining for named entity recognition tasks. |
| Outcome: | The results show that models can re-acquire pruned concepts after pruning . they also show that they can blend old and new concepts in individual neurons . |
Think While You Write: Hypothesis Verification Promotes Faithful Knowledge-to-Text Generation (2024.findings-naacl)
Copied to clipboard
| Challenge: | Knowledge-to-text generators often struggle to faithfully generate descriptions for input facts . we propose a decoding-only method to reduce hallucinations . |
| Approach: | They propose a decoding-only method to generate accurate descriptions for input facts . they use a Natural Language Inference model as the model and replace it with a task-specific HVM . |
| Outcome: | The proposed method improves faithfulness with minimal impact on quality and in/out-of-distribution evaluations. |
Can Large Language Models Follow Concept Annotation Guidelines? A Case Study on Scientific and Financial Domains (2024.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) are known to distill knowledge from vast datasets during the pretraining phase. |
| Approach: | They propose to use in-context concept definitions as prompts for zero-shot sentence labeling tasks. |
| Outcome: | The proposed model outperforms the leading open-source models and their proprietary APIs in most cases. |
PMIndiaSum: Multilingual and Cross-lingual Headline Summarization for Languages in India (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing datasets for Indian languages are limited in terms of coverage and size. |
| Approach: | They propose a multilingual and massively parallel summarization corpus focused on languages in India that provides a training and testing ground for four language families, 14 languages, and the largest to date with 196 language pairs. |
| Outcome: | The proposed dataset provides a training and testing ground for four language families, 14 languages, and the largest to date with 196 language pairs. |
Are Large Language Model Temporally Grounded? (2024.naacl-long)
Copied to clipboard
| Challenge: | Recent large language models lack a consistent temporal model of textual narratives . sentence ordering in unlabelled texts is only weakly correlated with event ordering . |
| Approach: | They evaluate LLMs with textual narratives and evaluate their common-sense knowledge . they find that LLM models struggle the most with self-consistency . |
| Outcome: | The proposed models lack a consistent temporal model of textual narratives. |