Knowing Your Nonlinearities: Shapley Interactions Reveal the Underlying Structure of Data
- URL: http://arxiv.org/abs/2403.13106v1
- Date: Tue, 19 Mar 2024 19:13:22 GMT
- Title: Knowing Your Nonlinearities: Shapley Interactions Reveal the Underlying Structure of Data
- Authors: Divyansh Singhvi, Andrej Erkelens, Raghav Jain, Diganta Misra, Naomi Saphra,
- Abstract summary: We use Shapley Taylor interaction indices (STII) to analyze the impact of underlying data structure on model representations.
Considering linguistic structure in masked and auto-regressive language models (ML and ALMs), we find that STII increases within idiomatic expressions.
Our speech model findings reflect the phonetic principal that the openness of the oral cavity determines how much a phoneme varies based on its context.
- Score: 8.029715695737567
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Measuring nonlinear feature interaction is an established approach to understanding complex patterns of attribution in many models. In this paper, we use Shapley Taylor interaction indices (STII) to analyze the impact of underlying data structure on model representations in a variety of modalities, tasks, and architectures. Considering linguistic structure in masked and auto-regressive language models (MLMs and ALMs), we find that STII increases within idiomatic expressions and that MLMs scale STII with syntactic distance, relying more on syntax in their nonlinear structure than ALMs do. Our speech model findings reflect the phonetic principal that the openness of the oral cavity determines how much a phoneme varies based on its context. Finally, we study image classifiers and illustrate that feature interactions intuitively reflect object boundaries. Our wide range of results illustrates the benefits of interdisciplinary work and domain expertise in interpretability research.
Related papers
- SALAD: Improving Robustness and Generalization through Contrastive Learning with Structure-Aware and LLM-Driven Augmented Data [15.366930934639838]
We propose SALAD, a novel approach to enhance model robustness and generalization.
Our method generates structure-aware and counterfactually augmented data for contrastive learning.
We validate our approach through experiments on three tasks: Sentiment Classification, Sexism Detection, and Natural Language Inference.
arXiv Detail & Related papers (2025-04-16T15:40:10Z) - A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models [74.48084001058672]
The rise of foundation models has transformed machine learning research.
multimodal foundation models (MMFMs) pose unique interpretability challenges beyond unimodal frameworks.
This survey explores two key aspects: (1) the adaptation of LLM interpretability methods to multimodal models and (2) understanding the mechanistic differences between unimodal language models and crossmodal systems.
arXiv Detail & Related papers (2025-02-22T20:55:26Z) - Analysis and Visualization of Linguistic Structures in Large Language Models: Neural Representations of Verb-Particle Constructions in BERT [0.0]
This study investigates the internal representations of verb-particle combinations within large language models (LLMs)
We analyse the representational efficacy of its layers for various verb-particle constructions such as 'agree on', 'come back', and 'give up'
Results show that BERT's middle layers most effectively capture syntactic structures, with significant variability in representational accuracy across different verb categories.
arXiv Detail & Related papers (2024-12-19T09:21:39Z) - Interpreting token compositionality in LLMs: A robustness analysis [10.777646083061395]
Constituent-Aware Pooling (CAP) is a methodology designed to analyse how large language models process linguistic structures.
CAP intervenes in model activations through constituent-based pooling at various model levels.
arXiv Detail & Related papers (2024-10-16T18:10:50Z) - Unsupervised Morphological Tree Tokenizer [36.584680344291556]
We introduce morphological structure guidance to tokenization and propose a deep model to induce character-level structures of words.<n>Our method is capable of inducing character-level structures that align with morphological rules without annotated training data.<n> Empirical results indicate that the proposed method effectively retains complete morphemes and outperforms widely adopted methods such as BPE and WordPiece.
arXiv Detail & Related papers (2024-06-21T15:35:49Z) - Explaining Text Similarity in Transformer Models [52.571158418102584]
Recent advances in explainable AI have made it possible to mitigate limitations by leveraging improved explanations for Transformers.
We use BiLRP, an extension developed for computing second-order explanations in bilinear similarity models, to investigate which feature interactions drive similarity in NLP models.
Our findings contribute to a deeper understanding of different semantic similarity tasks and models, highlighting how novel explainable AI methods enable in-depth analyses and corpus-level insights.
arXiv Detail & Related papers (2024-05-10T17:11:31Z) - Syntax-Informed Interactive Model for Comprehensive Aspect-Based
Sentiment Analysis [0.0]
We introduce an innovative model: Syntactic Dependency Enhanced Multi-Task Interaction Architecture (SDEMTIA) for comprehensive ABSA.
Our approach innovatively exploits syntactic knowledge (dependency relations and types) using a specialized Syntactic Dependency Embedded Interactive Network (SDEIN)
We also incorporate a novel and efficient message-passing mechanism within a multi-task learning framework to bolster learning efficacy.
arXiv Detail & Related papers (2023-11-28T16:03:22Z) - FLIP: Fine-grained Alignment between ID-based Models and Pretrained Language Models for CTR Prediction [49.510163437116645]
Click-through rate (CTR) prediction plays as a core function module in personalized online services.
Traditional ID-based models for CTR prediction take as inputs the one-hot encoded ID features of tabular modality.
Pretrained Language Models(PLMs) has given rise to another paradigm, which takes as inputs the sentences of textual modality.
We propose to conduct Fine-grained feature-level ALignment between ID-based Models and Pretrained Language Models(FLIP) for CTR prediction.
arXiv Detail & Related papers (2023-10-30T11:25:03Z) - Improving Open Information Extraction with Large Language Models: A
Study on Demonstration Uncertainty [52.72790059506241]
Open Information Extraction (OIE) task aims at extracting structured facts from unstructured text.
Despite the potential of large language models (LLMs) like ChatGPT as a general task solver, they lag behind state-of-the-art (supervised) methods in OIE tasks.
arXiv Detail & Related papers (2023-09-07T01:35:24Z) - Feature Interactions Reveal Linguistic Structure in Language Models [2.0178765779788495]
We study feature interactions in the context of feature attribution methods for post-hoc interpretability.
We work out a grey box methodology, in which we train models to perfection on a formal language classification task.
We show that under specific configurations, some methods are indeed able to uncover the grammatical rules acquired by a model.
arXiv Detail & Related papers (2023-06-21T11:24:41Z) - Competence-Based Analysis of Language Models [21.43498764977656]
CALM (Competence-based Analysis of Language Models) is designed to investigate LLM competence in the context of specific tasks.
We develop a new approach for performing causal probing interventions using gradient-based adversarial attacks.
We carry out a case study of CALM using these interventions to analyze and compare LLM competence across a variety of lexical inference tasks.
arXiv Detail & Related papers (2023-03-01T08:53:36Z) - Autoregressive Structured Prediction with Language Models [73.11519625765301]
We describe an approach to model structures as sequences of actions in an autoregressive manner with PLMs.
Our approach achieves the new state-of-the-art on all the structured prediction tasks we looked at.
arXiv Detail & Related papers (2022-10-26T13:27:26Z) - Dynamic Latent Separation for Deep Learning [67.62190501599176]
A core problem in machine learning is to learn expressive latent variables for model prediction on complex data.
Here, we develop an approach that improves expressiveness, provides partial interpretation, and is not restricted to specific applications.
arXiv Detail & Related papers (2022-10-07T17:56:53Z) - The Geometry of Self-supervised Learning Models and its Impact on
Transfer Learning [62.601681746034956]
Self-supervised learning (SSL) has emerged as a desirable paradigm in computer vision.
We propose a data-driven geometric strategy to analyze different SSL models using local neighborhoods in the feature space induced by each.
arXiv Detail & Related papers (2022-09-18T18:15:38Z) - Unsupervised Mismatch Localization in Cross-Modal Sequential Data [5.932046800902776]
We develop an unsupervised learning algorithm that can infer the relationship between content-mismatched cross-modal data.
We propose a hierarchical Bayesian deep learning model, named mismatch localization variational autoencoder (ML-VAE), that decomposes the generative process of the speech into hierarchically structured latent variables.
Our experimental results show that ML-VAE successfully locates the mismatch between text and speech, without the need for human annotations.
arXiv Detail & Related papers (2022-05-05T14:23:27Z) - Did the Cat Drink the Coffee? Challenging Transformers with Generalized
Event Knowledge [59.22170796793179]
Transformers Language Models (TLMs) were tested on a benchmark for the textitdynamic estimation of thematic fit
Our results show that TLMs can reach performances that are comparable to those achieved by SDM.
However, additional analysis consistently suggests that TLMs do not capture important aspects of event knowledge.
arXiv Detail & Related papers (2021-07-22T20:52:26Z) - Explaining Neural Network Predictions on Sentence Pairs via Learning
Word-Group Masks [21.16662651409811]
We propose the Group Mask (GMASK) method to implicitly detect word correlations by grouping correlated words from the input text pair together.
The proposed method is evaluated with two different model architectures (decomposable attention model and BERT) across four datasets.
arXiv Detail & Related papers (2021-04-09T17:14:34Z) - Structured Attention for Unsupervised Dialogue Structure Induction [110.12561786644122]
We propose to incorporate structured attention layers into a Variational Recurrent Neural Network (VRNN) model with discrete latent states to learn dialogue structure in an unsupervised fashion.
Compared to a vanilla VRNN, structured attention enables a model to focus on different parts of the source sentence embeddings while enforcing a structural inductive bias.
arXiv Detail & Related papers (2020-09-17T23:07:03Z) - Semantic Complexity in End-to-End Spoken Language Understanding [20.184305170102082]
We analyze the relationship between the performance of STI models and the difficulty of the use case to which they are applied.
We show that near-perfect performance metrics for STI models reported in the literature were obtained with datasets with low semantic complexity values.
arXiv Detail & Related papers (2020-08-06T20:18:53Z) - Exploiting Syntactic Structure for Better Language Modeling: A Syntactic
Distance Approach [78.77265671634454]
We make use of a multi-task objective, i.e., the models simultaneously predict words as well as ground truth parse trees in a form called "syntactic distances"
Experimental results on the Penn Treebank and Chinese Treebank datasets show that when ground truth parse trees are provided as additional training signals, the model is able to achieve lower perplexity and induce trees with better quality.
arXiv Detail & Related papers (2020-05-12T15:35:00Z) - Improve Variational Autoencoder for Text Generationwith Discrete Latent
Bottleneck [52.08901549360262]
Variational autoencoders (VAEs) are essential tools in end-to-end representation learning.
VAEs tend to ignore latent variables with a strong auto-regressive decoder.
We propose a principled approach to enforce an implicit latent feature matching in a more compact latent space.
arXiv Detail & Related papers (2020-04-22T14:41:37Z) - A Dependency Syntactic Knowledge Augmented Interactive Architecture for
End-to-End Aspect-based Sentiment Analysis [73.74885246830611]
We propose a novel dependency syntactic knowledge augmented interactive architecture with multi-task learning for end-to-end ABSA.
This model is capable of fully exploiting the syntactic knowledge (dependency relations and types) by leveraging a well-designed Dependency Relation Embedded Graph Convolutional Network (DreGcn)
Extensive experimental results on three benchmark datasets demonstrate the effectiveness of our approach.
arXiv Detail & Related papers (2020-04-04T14:59:32Z) - Temporal Embeddings and Transformer Models for Narrative Text
Understanding [72.88083067388155]
We present two approaches to narrative text understanding for character relationship modelling.
The temporal evolution of these relations is described by dynamic word embeddings, that are designed to learn semantic changes over time.
A supervised learning approach based on the state-of-the-art transformer model BERT is used instead to detect static relations between characters.
arXiv Detail & Related papers (2020-03-19T14:23:12Z) - An enhanced Tree-LSTM architecture for sentence semantic modeling using
typed dependencies [0.0]
Tree-based Long short term memory (LSTM) network has become state-of-the-art for modeling the meaning of language texts.
This paper proposes an enhanced LSTM architecture, called relation gated LSTM, which can model the relationship between two inputs of a sequence.
We also introduce a Tree-LSTM model called Typed Dependency Tree-LSTM that uses the sentence dependency parse structure and the dependency type to embed sentence meaning into a dense vector.
arXiv Detail & Related papers (2020-02-18T18:10:03Z)
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.