Learn from your own latents and not from tokens: A sample-complexity theory

• URL: http://arxiv.org/abs/2605.27734v1
• Date: Tue, 26 May 2026 22:16:42 GMT
• Title: Learn from your own latents and not from tokens: A sample-complexity theory
• Abstract summary: We study networks trained to predict their emphown latent representations of related views or masked regions.<n>We show that latent prediction achieves this with a number of samples emphconstant in $L$, up to logarithmic factors.<n>This suggests that explicit stacking such as H-JEPA is largely redundant.
• Score: 53.5821824211418
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generative models, from diffusion models to large language models, achieve remarkable performance but at a cost in training data orders of magnitude larger than what biological learners require. An alternative paradigm has emerged in which networks are trained to predict their \emph{own} latent representations of related views or masked regions, as in data2vec and JEPA -- an idea related to predictive-coding accounts of the cortex. Despite strong empirical results, the theoretical understanding of these methods remains limited. Central questions include: by how much does latent prediction actually improve data efficiency? Is there a benefit to stacking such methods into multi-scale hierarchies? We answer both using as data a tractable probabilistic context-free grammar that captures the compositional structure of natural language and images. Such a grammar generates strings of visible tokens by recursively applying production rules along a tree of hidden symbols of depth $L$. For such data, supervised or token-level SSL require a number of samples \emph{exponential} in $L$ to recover the latent tree; we prove that latent prediction achieves this with a number of samples \emph{constant} in $L$, up to logarithmic factors. We confirm this bound with (i) a hierarchical clustering algorithm, (ii) an end-to-end neural network whose predictor-clusterer modules predict their own latents at each level via gradient descent, and (iii) the first sample-complexity analysis of data2vec, which we show implicitly performs hierarchical latent prediction. This suggests that explicit stacking such as H-JEPA is largely redundant.

Related papers

• A Hierarchical Language Model with Predictable Scaling Laws and Provable Benefits of Reasoning [15.789193816051911]
  We introduce a family of synthetic languages with hierarchical structure.<n>The role of context length and reasoning in autoregressive generation can be analyzed precisely.
  arXiv  Detail & Related papers  (2026-05-13T15:42:26Z)
• Context-level Language Modeling by Learning Predictive Context Embeddings [79.00607069677393]
  We introduce textbfContextLM, a framework that augments standard pretraining with an inherent textbfnext-context prediction objective.<n>This mechanism trains the model to learn predictive representations of multi-token contexts, leveraging error signals derived from future token chunks.<n>Experiments on the GPT2 and Pythia model families, scaled up to $1.5$B parameters, show that ContextLM delivers consistent improvements in both perplexity and downstream task performance.
  arXiv  Detail & Related papers  (2025-10-23T07:09:45Z)
• How Do LLMs Use Their Depth? [17.148445769990907]
  We show that large language models do not use their depth uniformly, yet we still lack a fine-grained understanding of their layer-wise prediction dynamics.<n>We propose a "Guess-then-Refine" framework that explains how LLMs internally structure their computations to make predictions.
  arXiv  Detail & Related papers  (2025-10-21T17:59:05Z)
• Random-Set Large Language Models [4.308457163593758]
  Large Language Models (LLMs) are known to produce very high-quality tests and responses to our queries.<n>But how much can we trust this generated text?<n>We propose a novel Random-Set Large Language Model (RSLLM) approach which predicts finite random sets (belief functions) over the token space.
  arXiv  Detail & Related papers  (2025-04-25T05:25:27Z)
• I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data? [76.15163242945813]
  Large language models (LLMs) have led many to conclude that they exhibit a form of intelligence.<n>We introduce a novel generative model that generates tokens on the basis of human-interpretable concepts represented as latent discrete variables.
  arXiv  Detail & Related papers  (2025-03-12T01:21:17Z)
• TokenUnify: Scaling Up Autoregressive Pretraining for Neuron Segmentation [65.65530016765615]
  We propose a hierarchical predictive coding framework that captures multi-scale dependencies through three complementary learning objectives.<n> TokenUnify integrates random token prediction, next-token prediction, and next-all token prediction to create a comprehensive representational space.<n>We also introduce a large-scale EM dataset with 1.2 billion annotated voxels, offering ideal long-sequence visual data with spatial continuity.
  arXiv  Detail & Related papers  (2024-05-27T05:45:51Z)
• Exploring Beyond Logits: Hierarchical Dynamic Labeling Based on Embeddings for Semi-Supervised Classification [49.09505771145326]
  We propose a Hierarchical Dynamic Labeling (HDL) algorithm that does not depend on model predictions and utilizes image embeddings to generate sample labels. Our approach has the potential to change the paradigm of pseudo-label generation in semi-supervised learning.
  arXiv  Detail & Related papers  (2024-04-26T06:00:27Z)
• Clue Me In: Semi-Supervised FGVC with Out-of-Distribution Data [44.90231337626545]
  We propose a novel design specifically aimed at making out-of-distribution data work for semi-supervised visual classification. Our experimental results reveal that (i) the proposed method yields good robustness against out-of-distribution data, and (ii) it can be equipped with prior arts, boosting their performance.
  arXiv  Detail & Related papers  (2021-12-06T07:22:10Z)
• Complex Event Forecasting with Prediction Suffix Trees: Extended Technical Report [70.7321040534471]
  Complex Event Recognition (CER) systems have become popular in the past two decades due to their ability to "instantly" detect patterns on real-time streams of events. There is a lack of methods for forecasting when a pattern might occur before such an occurrence is actually detected by a CER engine. We present a formal framework that attempts to address the issue of Complex Event Forecasting.
  arXiv  Detail & Related papers  (2021-09-01T09:52:31Z)
• Understanding Self-supervised Learning with Dual Deep Networks [74.92916579635336]
  We propose a novel framework to understand contrastive self-supervised learning (SSL) methods that employ dual pairs of deep ReLU networks. We prove that in each SGD update of SimCLR with various loss functions, the weights at each layer are updated by a emphcovariance operator. To further study what role the covariance operator plays and which features are learned in such a process, we model data generation and augmentation processes through a emphhierarchical latent tree model (HLTM)
  arXiv  Detail & Related papers  (2020-10-01T17:51:49Z)

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.