Sometimes I am a Tree: Data Drives Unstable Hierarchical Generalization

• URL: http://arxiv.org/abs/2412.04619v3
• Date: Thu, 19 Dec 2024 17:51:34 GMT
• Title: Sometimes I am a Tree: Data Drives Unstable Hierarchical Generalization
• Authors: Tian Qin, Naomi Saphra, David Alvarez-Melis,
• Abstract summary: We use case studies of English grammar to explore how complex, diverse training data drives models to generalize OOD. We show that these factors are nuanced and that intermediate levels of diversity and complexity lead to inconsistent behavior across random seeds. Our findings emphasize the critical role of training data in shaping generalization patterns and illuminate how competing model strategies lead to inconsistent generalization outcomes across random seeds.
• Score: 15.028208772567487
• License:
• Abstract: Language models (LMs), like other neural networks, often favor shortcut heuristics based on surface-level patterns. Although LMs behave like n-gram models early in training, they must eventually learn hierarchical syntactic representations to correctly apply grammatical rules out-of-distribution (OOD). In this work, we use case studies of English grammar to explore how complex, diverse training data drives models to generalize OOD. We construct a framework that unifies our understanding of random variation with training dynamics, rule selection with memorization, and data diversity with complexity. We show that these factors are nuanced, and that intermediate levels of diversity and complexity lead to inconsistent behavior across random seeds and to unstable training dynamics. Our findings emphasize the critical role of training data in shaping generalization patterns and illuminate how competing model strategies lead to inconsistent generalization outcomes across random seeds. Code is available at https://github.com/sunnytqin/concept_comp.git.

Related papers

• How compositional generalization and creativity improve as diffusion models are trained [82.08869888944324]
  How many samples do generative models need to learn the composition rules, so as to produce a number of novel data? We consider diffusion models trained on simple context-free grammars - tree-like graphical models used to represent the structure of data such as language and images. We demonstrate that diffusion models learn compositional rules with the sample complexity required for clustering features with statistically similar context, a process similar to the word2vec.
  arXiv  Detail & Related papers  (2025-02-17T18:06:33Z)
• Exploring the Learning Capabilities of Language Models using LEVERWORLDS [23.40759867281453]
  Learning a model of a setting often involves learning both general structure rules and specific properties of the instance. This paper investigates the interplay between learning the general and the specific in various learning methods, with emphasis on sample efficiency.
  arXiv  Detail & Related papers  (2024-10-01T09:02:13Z)
• Learning Syntax Without Planting Trees: Understanding When and Why Transformers Generalize Hierarchically [74.96551626420188]
  Transformers trained on natural language data have been shown to learn its hierarchical structure and generalize to sentences with unseen syntactic structures. We investigate sources of inductive bias in transformer models and their training that could cause such generalization behavior to emerge.
  arXiv  Detail & Related papers  (2024-04-25T07:10:29Z)
• How to Plant Trees in Language Models: Data and Architectural Effects on the Emergence of Syntactic Inductive Biases [28.58785395946639]
  We show that pre-training can teach language models to rely on hierarchical syntactic features when performing tasks after fine-tuning. We focus on architectural features (depth, width, and number of parameters), as well as the genre and size of the pre-training corpus.
  arXiv  Detail & Related papers  (2023-05-31T14:38:14Z)
• On the Compositional Generalization Gap of In-Context Learning [73.09193595292233]
  We look at the gap between the in-distribution (ID) and out-of-distribution (OOD) performance of such models in semantic parsing tasks with in-context learning. We evaluate four model families, OPT, BLOOM, CodeGen and Codex on three semantic parsing datasets.
  arXiv  Detail & Related papers  (2022-11-15T19:56:37Z)
• Sequence-to-Sequence Learning with Latent Neural Grammars [12.624691611049341]
  Sequence-to-sequence learning with neural networks has become the de facto standard for sequence prediction tasks. While flexible and performant, these models often require large datasets for training and can fail spectacularly on benchmarks designed to test for compositional generalization. This work explores an alternative, hierarchical approach to sequence-to-sequence learning with quasi-synchronous grammars.
  arXiv  Detail & Related papers  (2021-09-02T17:58:08Z)
• Meta-Learning to Compositionally Generalize [34.656819307701156]
  We implement a meta-learning augmented version of supervised learning. We construct pairs of tasks for meta-learning by sub-sampling existing training data. Experimental results on the COGS and SCAN datasets show that our similarity-driven meta-learning can improve generalization performance.
  arXiv  Detail & Related papers  (2021-06-08T11:21:48Z)
• Structured Reordering for Modeling Latent Alignments in Sequence Transduction [86.94309120789396]
  We present an efficient dynamic programming algorithm performing exact marginal inference of separable permutations. The resulting seq2seq model exhibits better systematic generalization than standard models on synthetic problems and NLP tasks.
  arXiv  Detail & Related papers  (2021-06-06T21:53:54Z)
• Evading the Simplicity Bias: Training a Diverse Set of Models Discovers Solutions with Superior OOD Generalization [93.8373619657239]
  Neural networks trained with SGD were recently shown to rely preferentially on linearly-predictive features. This simplicity bias can explain their lack of robustness out of distribution (OOD) We demonstrate that the simplicity bias can be mitigated and OOD generalization improved.
  arXiv  Detail & Related papers  (2021-05-12T12:12:24Z)
• Neural Complexity Measures [96.06344259626127]
  We propose Neural Complexity (NC), a meta-learning framework for predicting generalization. Our model learns a scalar complexity measure through interactions with many heterogeneous tasks in a data-driven way.
  arXiv  Detail & Related papers  (2020-08-07T02:12:10Z)

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.