Mask-reconstruction pretraining helps in downstream tasks by capturing more semantic features.
problem How mask-reconstruction pretraining helps in downstream tasks and why it surpasses supervised learning.
method Theoretical analysis and experimental validation of mask-reconstruction pretraining (MRP) on auto-encoders.
result MRP provably captures more semantic features than supervised learning, leading to better performance in downstream tasks.
WeatherFormer learns robust weather features from small datasets.
problem Modeling complex weather dynamics from limited data.
method Pretrained transformer encoder on large satellite dataset, with spatiotemporal encoding.
result State-of-the-art performance in county-level soybean yield prediction and influenza forecasting.
CERT improves language understanding by contrastively learning sentence-level semantics.
problem Lack of sentence-level semantics in existing pretraining tasks.
method Contrastive self-supervised learning at the sentence level using back-translation augmentations.
result CERT outperforms BERT on 7 out of 11 GLUE benchmark tasks, achieving the same performance as BERT on 2 tasks.
Pretraining method enhances dialogue representation learning across various tasks.
problem Scarce labeled data for specific dialogue tasks.
method Multi-task unsupervised pretraining with natural training objectives.
result Significant improvement in downstream tasks without encoder discrimination.
Unsupervised algorithm parses CSG images into CFG without pretraining.
problem Sparse reward problem in unsupervised program synthesis for images.
method Grammar-encoded tree LSTM, entropy regularization, sampling without replacement.
result Recover meaningful programs in large search spaces (up to 3.8imes1028). Framework adds invariance to pretrained networks without fine-tuning.
problem Adding invariance to pretrained networks without altering original behavior.
method Post-training augmentation invariance framework with Markov-Wasserstein minimization and Wasserstein correlation maximization losses.
result Adapter networks improve classification accuracy on rotated and noisy images.
CFA improves model's ability to generalize across unseen domain-class combinations.
problem Challenges in real-world machine learning applications due to data distribution shifts and limited training data.
method Developed Compositional Feature Alignment (CFA) technique to improve CG ability of pretrained models.
result CFA outperforms common finetuning techniques in compositional generalization.
Study reveals class disparities in balanced datasets through spectral imbalance.
problem Class disparities in balanced datasets are overlooked despite model performance gaps.
method Developed a theoretical framework and studied 11 encoders to diagnose spectral imbalance.
result Identified spectral imbalance as a source of class disparities in balanced datasets.
RAE improves image representation learning with simplified design choices.
problem Improving image representation learning using pretrained vision encoders.
method Generalized RAE formulation, complementary working mechanisms of RAE and REPA, and free CFG guidance.
result RAEv2 achieves state-of-the-art results with 10x faster convergence and less training time.
This paper improves model generalization by integrating diverse pretrained models.
problem Leveraging diverse pretrained models for robust out-of-distribution generalization.
method Characterize and integrate diverse pretrained models based on diversity and correlation shifts.
result Demonstrates state-of-the-art out-of-distribution generalization performance.
New study finds best language model architecture and pretraining objective for zero-shot tasks.
problem Evaluating which language model architectures and pretraining objectives best enable zero-shot generalization.
method Compared three model architectures and two pretraining objectives across 170 billion tokens, with and without finetuning.
result Causal decoder-only models trained on autoregressive language modeling exhibit strongest zero-shot generalization.
The dissertation establishes a contexture theory to mathematically characterize representation learning.
problem The lack of a scientific understanding of representation learning in foundation models.
method Introduces the contexture theory as a unified framework for analyzing representation learning methods.
result Representation learning is optimal when the association between input and context is neither too strong nor too weak.
Two CSSL-based methods improve graph classification with limited labeled data.
problem Limited labeled data for graph classification leads to overfitting.
method Contrastive self-supervised learning (CSSL) for graph encoders pretraining and regularization.
result CSSL methods reduce overfitting and improve graph classification accuracy.
Unified framework for ICL in causal and masked models.
problem Understanding ICL in masked language models and comparing it to causal models.
method Developed a statistical learning framework representing context by empirical measure and predicting using context and query.
result Upper bounds for masked and autoregressive objectives under Wasserstein-type regularity conditions.
We present BART, a denoising autoencoder for pretraining sequence-to-sequence models. BART is trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text. It uses a standard Tranformer-based neural machine translation architecture which, despite its simpl…
Researchers propose a new SSL risk decomposition method to evaluate and improve self-supervised learning models.
problem Self-supervised learning evaluation is limited to a single metric, providing little insight into model performance and improvement.
method Proposes an SSL risk decomposition that considers four error components: approximation, representation usability, probe generalization, and encoder generalization.
result Analysis of 169 SSL vision models reveals the main sources of error and provides insights for improving SSL models in specific settings.
POLAR learns efficient data acquisition policies using pretrained belief representations.
problem Challenges in learning effective policies for adaptive data acquisition.
method POLAR decouples representation learning from policy learning by leveraging pretrained predictive foundation models as belief-state encoders.
result POLAR outperforms state-of-the-art methods across diverse tasks while requiring fewer training samples.
A new approach to protein language models combines latent space prediction with masked language modeling.
problem Improving protein language models by predicting amino acid identities at masked positions.
method A variant of masked language modeling that predicts latent targets only at masked positions, retaining the MLM cross-entropy.
result The new approach outperforms pure masked language modeling on 11 out of 16 downstream tasks.
Transformers learn new tasks from few examples via optimal approximation.
problem Learning new tasks from limited examples using large language models.
method Developed approximation and generalization error bounds for transformers trained on nonparametric regression tasks.
result Transformers achieve minimax optimal estimation risk in context.
Transformer model pretrains on synthetic graphs for AD detection.
problem Limited labeled data and class imbalance in AD diagnosis.
method Diffusion-generated synthetic graphs, Graph Transformers, transfer learning.
result Framework outperforms baselines in AD diagnosis metrics.
Improves retrieval accuracy for hierarchical documents, especially for distant matches.
problem Limited expressive power of dual encoder models in hierarchical retrieval.
method Proves feasibility of DEs for HR, introduces pretrain-finetune recipe to improve long-distance retrieval.
result Pretrain-finetune boosts recall on long-distance pairs from 19% to 76%.
Quantum circuit optimization speeds up financial derivatives pricing.
problem Efficiently pricing financial derivatives on quantum computers.
method Pretraining conditional parameterized circuits for state-dependent functions.
result Quantum circuit implementation of derivatives' payoff function is more efficient.
Adversarially trained generative models (GANs) have recently achieved compelling image synthesis results. But despite early successes in using GANs for unsupervised representation learning, they have since been superseded by approaches based on self-supervision. In this work we show that progress in image generation qu…
Framework identifies population quantities from MNAR feedback using weak shadow variables from pretrained models.
problem Estimating mean outcomes from MNAR user feedback with bias and lack of identification.
method Develops a partial identification framework using linear programs and weak shadow variables from pretrained models.
result Bounds on estimand are obtained by solving linear programs incorporating pretrained model predictions.
This work analyzes CoT prompting methods from a statistical estimation perspective.
problem Improving the effectiveness of LLMs in solving multi-step reasoning problems.
method Introducing a multi-step latent variable model to characterize CoT prompting from a statistical estimation viewpoint.
result The CoT estimator is equivalent to a Bayesian estimator when the pretraining dataset is large.
We introduce and study the problem of Online Continual Compression, where one attempts to simultaneously learn to compress and store a representative dataset from a non i.i.d data stream, while only observing each sample once. A naive application of auto-encoders in this setting encounters a major challenge: representa…
We incorporate Tensor-Product Representations within the Transformer in order to better support the explicit representation of relation structure. Our Tensor-Product Transformer (TP-Transformer) sets a new state of the art on the recently-introduced Mathematics Dataset containing 56 categories of free-form math word-pr…
In this paper, we introduce the problem of jointly learning feed-forward neural networks across a set of relevant but diverse datasets. Compared to learning a separate network from each dataset in isolation, joint learning enables us to extract correlated information across multiple datasets to significantly improve th…
Task-agnostic data augmentation shows little benefit for pretrained transformers.
problem Evaluating the effectiveness of task-agnostic data augmentation on pretrained transformers.
method Conducted a systematic examination of two data augmentation techniques (Easy Data Augmentation and Back-Translation) across 5 tasks, 6 datasets, and 3 pretrained transformer models.
result Data augmentation techniques previously effective for non-pretrained models fail to consistently improve performance for pretrained transformers, even with limited training data.
Study evaluates how much knowledge LLMs have by comparing their prediction accuracy to flexible models.
problem Evaluating the predictive power of LLMs without access to their training data.
method Equivalent sample size measure, comparing LLM's prediction error to flexible models trained on varying amounts of domain-specific data.
result LLMs encode varying amounts of predictive information across different economic variables.
This paper investigates the influence of different acoustic features, audio-events based features and automatic speech translation based lexical features in complex emotion recognition such as curiosity. Pretrained networks, namely, AudioSet Net, VoxCeleb Net and Deep Speech Net trained extensively for different speech…
This paper proposes a set of new error criteria and learning approaches, Adaptive Normalized Risk-Averting Training (ANRAT), to attack the non-convex optimization problem in training deep neural networks (DNNs). Theoretically, we demonstrate its effectiveness on global and local convexity lower-bounded by the standard …
This study examines how large language models learn in-context and provides insights into their performance and architecture.
problem Understanding how large language models learn in-context and their performance metrics.
method Bayesian model averaging, parameterization, and statistical analysis of transformer architecture.
result Transformer architecture enables in-context learning through attention mechanisms and fine-grained statistical analysis.
ContextFlow++ improves generative models by conditioning on mixed-variable contexts.
problem Lack of effective methods for context conditioning in flow-based generative models.
method Proposes ContextFlow++ with additive conditioning and mixed-variable architecture.
result ContextFlow++ achieves higher performance metrics and faster training.
Pretraining models improves text classification accuracy, but diminishing returns are observed with large datasets.
problem Improving text classification accuracy with pretrained models.
method Examined the benefits of pretrained models on text classification tasks with varying amounts of training data.
result As the number of training examples grows into the millions, the accuracy gap between pretrained BERT-based models and vanilla LSTM narrows to within 1%.
New framework selects high-quality pretraining data without training LLMs.
problem Slow progress in understanding pretraining data due to costly experiments.
method Statistical framework based on perplexity-benchmark correlations.
result Approach outperforms existing methods on multiple benchmarks.
A method for self-supervised learning of multivariate time series data across varying channels.
problem Labeling multivariate biomedical time series data is laborious and expensive.
method Proposes a multi-view self-supervised learning approach using a message passing neural network to extract a single representation across channels.
result Our method, combined with the TS2Vec loss, outperforms all other methods in most settings.
Algorithm adapts pretrained semantic segmentation models to new domains.
problem Adapting pretrained models to new, unlabeled domains without source data.
method Learn prototypical distribution in embedding space, align target domain with source domain.
result Method achieves competitive performance on benchmark tasks.
Paper studies how few pretraining tasks are needed for a linear model to solve new tasks.
problem How many pretraining tasks are needed for a linear model to solve new tasks?
method Pretrained a linear attention model for linear regression with a Gaussian prior.
result Effective pretraining requires a small number of independent tasks, and the model closely matches Bayes optimal.
Fine-grained pretraining improves neural network's ability to learn rare features.
problem Improving generalization in deep learning models.
method Introducing a hierarchical multi-view structure to confine input data distribution.
result Fine-grained pretraining leads to better accuracy on hard downstream test samples.
Method constructs finance LLMs without instruction data using pretraining and model merging.
problem Developing domain-specific LLMs for finance is resource-intensive.
method Continual pretraining on financial data + model merging of instruction-tuned and domain-specific pretrained vectors.
result Successfully constructs instruction-tuned LLMs for finance without additional instruction data.
A self-supervised debiasing method using rank regularization mitigates spurious correlations in neural networks.
problem Spurious correlations cause biases in deep neural networks, affecting generalization.
method Spectral analysis of latent representations, rank regularization, self-supervised pretraining, debiasing of downstream tasks.
result The proposed framework significantly improves generalization performance and outperforms supervised debiasing approaches.
Variational Auto-Encoders have often been used for unsupervised pretraining, feature extraction and out-of-distribution and anomaly detection in the medical field. However, VAEs often lack the ability to produce sharp images and learn high-level features. We propose to alleviate these issues by adding a new branch to c…
Paper introduces a multi-stage influence function to track model predictions.
problem Improving natural language processing and computer vision performance.
method Develops a multi-stage influence function to track predictions from finetuned models back to pretraining data.
result Identifies pretraining examples contributing most to finetuning task predictions.
Synthetic continued pretraining enhances model performance with synthetic data.
problem Data inefficiency in pretrained models when adapting to domain-specific documents.
method Synthetic data augmentation using EntiGraph to create a large synthetic corpus.
result Language models can answer questions and follow instructions without access to domain-specific documents.
Theoretical analysis of data quality and synergies in LLMs.
problem Understanding why different training methods require different amounts of data.
method Theoretical analysis of transformers trained on a weight prediction task for linear regression.
result SFT excels on smaller datasets challenging for the pretrained model, while RL benefits from large, not overly difficult data.
Denoised smoothing defends pretrained classifiers against adversarial attacks.
problem Adversarial attacks on pretrained classifiers.
method Prepending a denoiser to any off-the-shelf classifier using randomized smoothing.
result Guaranteed ℓp-robustness to adversarial examples without modifying the pretrained classifier. Spiking neural networks (SNNs) are distributed trainable systems whose computing elements, or neurons, are characterized by internal analog dynamics and by digital and sparse synaptic communications. The sparsity of the synaptic spiking inputs and the corresponding event-driven nature of neural processing can be levera…