Conditional probing: measuring usable information beyond a baseline

• URL: http://arxiv.org/abs/2109.09234v1
• Date: Sun, 19 Sep 2021 21:56:58 GMT
• Title: Conditional probing: measuring usable information beyond a baseline
• Authors: John Hewitt, Kawin Ethayarajh, Percy Liang, Christopher D. Manning
• Abstract summary: One suggests that a representation encodes a property if probing that representation produces higher accuracy than probing a baseline representation. We propose conditional probing, which explicitly conditions on the information in the baseline. In a case study, we find that after conditioning on non-contextual word embeddings, properties like part-of-speech are accessible at deeper layers of a network.
• Score: 103.93673427217527
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Probing experiments investigate the extent to which neural representations make properties -- like part-of-speech -- predictable. One suggests that a representation encodes a property if probing that representation produces higher accuracy than probing a baseline representation like non-contextual word embeddings. Instead of using baselines as a point of comparison, we're interested in measuring information that is contained in the representation but not in the baseline. For example, current methods can detect when a representation is more useful than the word identity (a baseline) for predicting part-of-speech; however, they cannot detect when the representation is predictive of just the aspects of part-of-speech not explainable by the word identity. In this work, we extend a theory of usable information called $\mathcal{V}$-information and propose conditional probing, which explicitly conditions on the information in the baseline. In a case study, we find that after conditioning on non-contextual word embeddings, properties like part-of-speech are accessible at deeper layers of a network than previously thought.

Related papers

• Surprise! Uniform Information Density Isn't the Whole Story: Predicting Surprisal Contours in Long-form Discourse [54.08750245737734]
  We propose that speakers modulate information rate based on location within a hierarchically-structured model of discourse. We find that hierarchical predictors are significant predictors of a discourse's information contour and that deeply nested hierarchical predictors are more predictive than shallow ones.
  arXiv  Detail & Related papers  (2024-10-21T14:42:37Z)
• Analyzing Text Representations by Measuring Task Alignment [2.198430261120653]
  We develop a task alignment score based on hierarchical clustering that measures alignment at different levels of granularity. Our experiments on text classification validate our hypothesis by showing that task alignment can explain the classification performance of a given representation.
  arXiv  Detail & Related papers  (2023-05-31T11:20:48Z)
• Textual Entailment Recognition with Semantic Features from Empirical Text Representation [60.31047947815282]
  A text entails a hypothesis if and only if the true value of the hypothesis follows the text. In this paper, we propose a novel approach to identifying the textual entailment relationship between text and hypothesis. We employ an element-wise Manhattan distance vector-based feature that can identify the semantic entailment relationship between the text-hypothesis pair.
  arXiv  Detail & Related papers  (2022-10-18T10:03:51Z)
• Saliency Map Verbalization: Comparing Feature Importance Representations from Model-free and Instruction-based Methods [6.018950511093273]
  Saliency maps can explain a neural model's predictions by identifying important input features. We formalize the underexplored task of translating saliency maps into natural language. We compare two novel methods (search-based and instruction-based verbalizations) against conventional feature importance representations.
  arXiv  Detail & Related papers  (2022-10-13T17:48:15Z)
• Measuring the Interpretability of Unsupervised Representations via Quantized Reverse Probing [97.70862116338554]
  We investigate the problem of measuring interpretability of self-supervised representations. We formulate the latter as estimating the mutual information between the representation and a space of manually labelled concepts. We use our method to evaluate a large number of self-supervised representations, ranking them by interpretability.
  arXiv  Detail & Related papers  (2022-09-07T16:18:50Z)
• Fair Interpretable Learning via Correction Vectors [68.29997072804537]
  We propose a new framework for fair representation learning centered around the learning of "correction vectors" The corrections are then simply summed up to the original features, and can therefore be analyzed as an explicit penalty or bonus to each feature. We show experimentally that a fair representation learning problem constrained in such a way does not impact performance.
  arXiv  Detail & Related papers  (2022-01-17T10:59:33Z)
• Predicting What You Already Know Helps: Provable Self-Supervised Learning [60.27658820909876]
  Self-supervised representation learning solves auxiliary prediction tasks (known as pretext tasks) without requiring labeled data. We show a mechanism exploiting the statistical connections between certain em reconstruction-based pretext tasks that guarantee to learn a good representation. We prove the linear layer yields small approximation error even for complex ground truth function class.
  arXiv  Detail & Related papers  (2020-08-03T17:56:13Z)
• Measuring Memorization Effect in Word-Level Neural Networks Probing [0.9156064716689833]
  We propose a simple general method for measuring the memorization effect, based on a symmetric selection of test words seen versus unseen in training. Our method can be used to explicitly quantify the amount of memorization happening in a probing setup, so that an adequate setup can be chosen and the results of the probing can be interpreted with a reliability estimate.
  arXiv  Detail & Related papers  (2020-06-29T14:35:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.