Abhilasha Ravichander
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Abhilasha Ravichander has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Eduard Hovy
The University of Melbourne
3 shared publications Hinrich Schütze
Center for Information and Language Processing (CIS), LMU Munich, Germany
2 shared publications- Nora Kassner
Center for Information and Language Processing (CIS), LMU Munich, Germany
2 shared publications Shauli Ravfogel
Computer Science Department, Bar Ilan University, Israel
2 shared publications- Shruti Rijhwani
Language Technologies Institute Carnegie Mellon University
2 shared publications - Yanai Elazar
Allen Institute for AI
2 shared publications
Latest Publications
- Open Access
- 6 December 2021
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 9, 1407
Erratum: Measuring and Improving Consistency in Pretrained Language Models
During production of this paper, an error was introduced to the formula on the bottom of the right column of page 1020. In the last two terms of the formula, the n and m subscripts were swapped. The correct formula is: The paper has been updated.- 6 December 2021
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 9, 1407
Open AccessErratum: Measuring and Improving Consistency in Pretrained Language Models
During production of this paper, an error was introduced to the formula on the bottom of the right column of page 1020. In the last two terms of the formula, the n and m subscripts were swapped. The correct formula is: The paper has been updated. - Top Cited PapersOpen Access
- 6 December 2021
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 9, 1012-1031
Measuring and Improving Consistency in Pretrained Language Models
Consistency of a model—that is, the invariance of its behavior under meaning-preserving alternations in its input—is a highly desirable property in natural language processing. In this paper we study the question: Are Pretrained Language Models (PLMs) consistent with respect to factual knowledge? To this end, we create ParaRel🤘, a high-quality resource of cloze-style query English paraphrases. It contains a total of 328 paraphrases for 38 relations. Using ParaRel🤘, we show that the consistency of all PLMs we experiment with is poor— though with high variance between relations. Our analysis of the representational spaces of PLMs suggests that they have a poor structure and are currently not suitable for representing knowledge robustly. Finally, we propose a method for improving model consistency and experimentally demonstrate its effectiveness. 1
- 6 December 2021
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 9, 1012-1031
Top Cited PapersOpen AccessMeasuring and Improving Consistency in Pretrained Language Models
Consistency of a model—that is, the invariance of its behavior under meaning-preserving alternations in its input—is a highly desirable property in natural language processing. In this paper we study the question: Are Pretrained Language Models (PLMs) consistent with respect to factual knowledge? To this end, we create ParaRel🤘, a high-quality resource of cloze-style query English paraphrases. It contains a total of 328 paraphrases for 38 relations. Using ParaRel🤘, we show that the consistency of all PLMs we experiment with is poor— though with high variance between relations. Our analysis of the representational spaces of PLMs suggests that they have a poor structure and are currently not suitable for representing knowledge robustly. Finally, we propose a method for improving model consistency and experimentally demonstrate its effectiveness. 1
- Open Access
Probing the Probing Paradigm: Does Probing Accuracy Entail Task Relevance?
Although neural models have achieved impressive results on several NLP benchmarks, little is understood about the mechanisms they use to perform language tasks. Thus, much recent attention has been devoted to analyzing the sentence representations learned by neural encoders, through the lens of `probing' tasks. However, to what extent was the information encoded in sentence representations, as discovered through a probe, actually used by the model to perform its task? In this work, we examine this probing paradigm through a case study in Natural Language Inference, showing that models can learn to encode linguistic properties even if they are not needed for the task on which the model was trained. We further identify that pretrained word embeddings play a considerable role in encoding these properties rather than the training task itself, highlighting the importance of careful controls when designing probing experiments. Finally, through a set of controlled synthetic tasks, we demonstrate models can encode these properties considerably above chance-level even when distributed in the data as random noise, calling into question the interpretation of absolute claims on probing tasks.
Open AccessProbing the Probing Paradigm: Does Probing Accuracy Entail Task Relevance?
Although neural models have achieved impressive results on several NLP benchmarks, little is understood about the mechanisms they use to perform language tasks. Thus, much recent attention has been devoted to analyzing the sentence representations learned by neural encoders, through the lens of `probing' tasks. However, to what extent was the information encoded in sentence representations, as discovered through a probe, actually used by the model to perform its task? In this work, we examine this probing paradigm through a case study in Natural Language Inference, showing that models can learn to encode linguistic properties even if they are not needed for the task on which the model was trained. We further identify that pretrained word embeddings play a considerable role in encoding these properties rather than the training task itself, highlighting the importance of careful controls when designing probing experiments. Finally, through a set of controlled synthetic tasks, we demonstrate models can encode these properties considerably above chance-level even when distributed in the data as random noise, calling into question the interpretation of absolute claims on probing tasks.
- Open Access
Preserving Intermediate Objectives: One Simple Trick to Improve Learning for Hierarchical Models
Hierarchical models are utilized in a wide variety of problems which are characterized by task hierarchies, where predictions on smaller subtasks are useful for trying to predict a final task. Typically, neural networks are first trained for the subtasks, and the predictions of these networks are subsequently used as additional features when training a model and doing inference for a final task. In this work, we focus on improving learning for such hierarchical models and demonstrate our method on the task of speaker trait prediction. Speaker trait prediction aims to computationally identify which personality traits a speaker might be perceived to have, and has been of great interest to both the Artificial Intelligence and Social Science communities. Persuasiveness prediction in particular has been of interest, as persuasive speakers have a large amount of influence on our thoughts, opinions and beliefs. In this work, we examine how leveraging the relationship between related speaker traits in a hierarchical structure can help improve our ability to predict how persuasive a speaker is. We present a novel algorithm that allows us to backpropagate through this hierarchy. This hierarchical model achieves a 25% relative error reduction in classification accuracy over current state-of-the art methods on the publicly available POM dataset.
Open AccessPreserving Intermediate Objectives: One Simple Trick to Improve Learning for Hierarchical Models
Hierarchical models are utilized in a wide variety of problems which are characterized by task hierarchies, where predictions on smaller subtasks are useful for trying to predict a final task. Typically, neural networks are first trained for the subtasks, and the predictions of these networks are subsequently used as additional features when training a model and doing inference for a final task. In this work, we focus on improving learning for such hierarchical models and demonstrate our method on the task of speaker trait prediction. Speaker trait prediction aims to computationally identify which personality traits a speaker might be perceived to have, and has been of great interest to both the Artificial Intelligence and Social Science communities. Persuasiveness prediction in particular has been of interest, as persuasive speakers have a large amount of influence on our thoughts, opinions and beliefs. In this work, we examine how leveraging the relationship between related speaker traits in a hierarchical structure can help improve our ability to predict how persuasive a speaker is. We present a novel algorithm that allows us to backpropagate through this hierarchy. This hierarchical model achieves a 25% relative error reduction in classification accuracy over current state-of-the art methods on the publicly available POM dataset.
- Open Access
Does the Geometry of Word Embeddings Help Document Classification? A Case Study on Persistent Homology Based Representations
We investigate the pertinence of methods from algebraic topology for text data analysis. These methods enable the development of mathematically-principled isometric-invariant mappings from a set of vectors to a document embedding, which is stable with respect to the geometry of the document in the selected metric space. In this work, we evaluate the utility of these topology-based document representations in traditional NLP tasks, specifically document clustering and sentiment classification. We find that the embeddings do not benefit text analysis. In fact, performance is worse than simple techniques like $\textit{tf-idf}$, indicating that the geometry of the document does not provide enough variability for classification on the basis of topic or sentiment in the chosen datasets.Open AccessDoes the Geometry of Word Embeddings Help Document Classification? A Case Study on Persistent Homology Based Representations
We investigate the pertinence of methods from algebraic topology for text data analysis. These methods enable the development of mathematically-principled isometric-invariant mappings from a set of vectors to a document embedding, which is stable with respect to the geometry of the document in the selected metric space. In this work, we evaluate the utility of these topology-based document representations in traditional NLP tasks, specifically document clustering and sentiment classification. We find that the embeddings do not benefit text analysis. In fact, performance is worse than simple techniques like $\textit{tf-idf}$, indicating that the geometry of the document does not provide enough variability for classification on the basis of topic or sentiment in the chosen datasets.