Papers by Dan Friedman
Single-dataset Experts for Multi-dataset Question Answering (2021.emnlp-main)
Copied to clipboard
| Challenge: | Prior work has focused on training one network on multiple datasets to build a model that performs well on all of the training datasets and generalizes and transfers better to new datasets. |
| Approach: | They combine multiple reading comprehension datasets to build a multi-dataset question answering model with an ensemble of single-data set experts. |
| Outcome: | The proposed model outperforms baseline models in in-distribution accuracy and generalization and transfer performance. |
Automatic Argument Quality Assessment - New Datasets and Methods (D19-1)
Copied to clipboard
Assaf Toledo, Shai Gretz, Edo Cohen-Karlik, Roni Friedman, Elad Venezian, Dan Lahav, Michal Jacovi, Ranit Aharonov, Noam Slonim
| Challenge: | 6.3k arguments were collected from contributors of various levels, and are released as part of this work. |
| Approach: | They propose to use a language model to annotate arguments for argument ranking and argument-pair classification. |
| Outcome: | The proposed methods outperform state-of-the-art methods in the argument ranking task and argument-pair classification task. |
From Arguments to Key Points: Towards Automatic Argument Summarization (2020.acl-main)
Copied to clipboard
| Challenge: | Recent work on topic-related argument mining has made it difficult to read and digest large amounts of information. |
| Approach: | They propose to represent arguments as a small set of talking points, termed key points, each scored according to its salience. |
| Outcome: | The proposed method can predict key points in advance, and it performs well. |
Quantitative argument summarization and beyond: Cross-domain key point analysis (2020.emnlp-main)
Copied to clipboard
| Challenge: | Recent work on multi-document summarization lacks quantitative aspect of summarizing views, arguments or opinions . authors develop method for automatic extraction of key points, which is comparable to a human expert . |
| Approach: | They propose to map arguments to a small set of expert-generated key points . they demonstrate that the applicability of key point analysis goes well beyond argumentation data . |
| Outcome: | The proposed method outperforms arguments in municipal surveys and user reviews . it is shown that the extraction of key points is comparable to a human expert . |
Measuring Inductive Biases of In-Context Learning with Underspecified Demonstrations (2023.acl-long)
Copied to clipboard
| Challenge: | In-context learning is an important paradigm for adapting large language models to new tasks . but the generalization behavior of ICL remains poorly understood . |
| Approach: | They characterize the feature biases of large language models by constructing underspecified demonstrations . they find that LLMs exhibit clear feature bias, and they evaluate interventions . |
| Outcome: | The proposed model prefers the "default" task features over distractor features more often than the base model. |
Benchmark Data and Evaluation Framework for Intent Discovery Around COVID-19 Vaccine Hesitancy (2023.findings-eacl)
Copied to clipboard
Shai Gretz, Assaf Toledo, Roni Friedman, Dan Lahav, Rose Weeks, Naor Bar-Zeev, João Sedoc, Pooja Sangha, Yoav Katz, Noam Slonim
| Challenge: | As COVID-19 vaccines were rolled out, they were met with widespread hesitancy. |
| Approach: | They propose a new framework for intent discovery that leverages existing intent classifiers to provide a real-world conversational dataset of conversations conducted by actual users with VIRA. |
| Outcome: | The proposed framework enables users to find out what they are doing and why they are hesitant. |
Representing Rule-based Chatbots with Transformers (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing work on how Transformers can solve synthetic tasks has not explored how to extend this to a conversational setting. |
| Approach: | They propose to use ELIZA as a framework for formal mechanistic analysis of Transformers . they propose to model local pattern matching and long-term dialogue state tracking . |
| Outcome: | The proposed model can be extended to model key aspects of conversation, the authors show . their model favors an induction head mechanism over a more precise copying mechanism . |
Syntax-aware Neural Semantic Role Labeling with Supertags (N19-1)
Copied to clipboard
| Challenge: | a new syntax-aware model for dependency-based semantic role labeling outperforms syntax-based models for English and Spanish. |
| Approach: | They propose a syntax-aware model for dependency-based semantic role labeling that outperforms syntax-based models for English and Spanish. |
| Outcome: | The proposed model outperforms syntax-agnostic models for English and Spanish. |
Finding Dataset Shortcuts with Grammar Induction (2022.emnlp-main)
Copied to clipboard
| Challenge: | Prior work on shortcut detection focused on enumerating features like unigrams or bigrams . prior work relied on post-hoc models that reveal qualitative patterns without a clear statistical interpretation . |
| Approach: | They propose to use probabilistic grammars to characterize and discover shortcuts in NLP datasets using context-free grammars and synchronous context- free grammars. |
| Outcome: | The proposed grammars reveal interesting shortcut features in a number of datasets, including simple and high-level features, and automatically identify groups of test examples on which conventional classifiers fail. |
Factual Probing Is [MASK]: Learning vs. Learning to Recall (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing methods for factual probing can interpret the model’s prediction accuracy as a lower bound on the amount of factual information it encodes. |
| Approach: | They propose a method which directly optimizes in continuous embedding space and can predict an additional 6.4% of facts in the LAMA benchmark. |
| Outcome: | The proposed method outperforms the best previous prompt method by 6.4% on the LAMA benchmark. |