Audited Conformal Prediction improves conditional coverage in pretrained models under distribution shift.
problem Uncertainty quantification for pretrained models under unknown distribution shift
method Leverages a small labeled dataset to train an audit model for marginal coverage, integrates outputs into conformal prediction framework
result Significantly higher conditional coverage than existing approaches
Transformers can learn new tasks from diverse pretraining data but struggle with out-of-domain tasks.
problem Transformer models' ability to learn new tasks in-context is limited by their pretraining data coverage.
method Investigation of transformer models trained on ( x , f ( x ) ) (x, f(x)) ( x , f ( x )) pairs, comparing in-context learning capabilities across different task families. result Transformers can identify and learn within task families in their pretraining data but fail with out-of-domain tasks.
New research reveals how the pretraining distribution affects in-context learning in large language models.
problem Understanding how the pretraining distribution influences in-context learning in large language models.
method Developed a theoretical framework to characterize the relationship between pretraining distribution properties and in-context learning performance.
result Characterized a fundamental trade-off between robust task selection and generalization in ICL due to the pretraining distribution's statistical properties.
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.
CAOS aggregates multiple one-shot predictors for efficient uncertainty quantification.
problem Lack of principled uncertainty quantification in one-shot prediction.
method CAOS, a conformal framework that aggregates multiple one-shot predictors and uses a leave-one-out calibration scheme.
result CAOS produces smaller prediction sets with reliable coverage compared to split conformal baselines.
New framework analyzes LLM personalization trade-offs under congestion.
problem Tension between personalization and resource sharing in LLMs.
method Developed a statistical-economic framework to model user incentives.
result Congestion can flip rankings of SFT and ICL, and offers both methods never hurt profits.
New theory explains how strong models can learn from weak ones.
problem Learning from weak, incomplete, or incorrect labels.
method New bounds based on data distribution and student hypothesis class.
result Existing weak supervision theory fails to account for pseudolabel correction and coverage expansion.
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.
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.
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.
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.
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.
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%.
This study examines how the size and alignment of pretraining data affect the performance of large language models on downstream tasks.
problem Understanding how the size and alignment of pretraining data impact the performance of large language models on downstream tasks.
method Investigated the scaling behavior of large language models in a transfer learning setting, focusing on machine translation tasks.
result The size of the finetuning dataset and the distribution alignment between pretraining and downstream data significantly influence the scaling behavior of downstream performance.
Publicly pretraining models on Web data may undermine differential privacy.
problem The use of large Web-scraped datasets in differential privacy models.
method Critical review of leveraging pretrained models on public datasets for differential privacy.
result Publicizing pretrained models as 'private' could harm trust and generalize poorly.
Study shows pretraining and finetuning can effectively tackle covariate shift in linear regression.
problem Linear regression under covariate shift where source and target distributions differ but conditional distribution remains similar.
method Pretraining on source data and finetuning on target data using online SGD.
result Transfer learning with O ( N 2 ) O(N^2) O ( N 2 ) source data is as effective as supervised learning with N N N target data. Model combination, often regarded as a key sub-field of ensemble learning, has been widely used in both academic research and industry applications. To facilitate this process, we propose and implement an easy-to-use Python toolkit, combo, to aggregate models and scores under various scenarios, including classification…
This study compares transfer learning and self-supervised learning for better model performance.
problem Choosing between transfer learning and self-supervised learning for optimal model performance.
method Comprehensive comparative study of transfer learning and self-supervised learning under various data and task properties.
result Self-supervised learning outperforms transfer learning in certain applications, and vice versa.
Algorithm improves decision-making with partially observed contexts using pretrained models.
problem Improving decision-making with partially observed contexts in online linear contextual bandits.
method PULSE-UCB algorithm that uses pretrained models trained on auxiliary data to impute missing features.
result Achieves near-optimal performance in i.i.d. context case with Hölder-smooth missing features.
Transformers learn to make decisions in new contexts from offline data.
problem Understanding when and how transformers can perform in-context reinforcement learning.
method Theoretical framework analyzing supervised pretraining for ICRL, including algorithm distillation and decision-pretrained transformers.
result Transformers can efficiently approximate optimal reinforcement learning algorithms for various environments.
Transformers fine-tuned on synthetic data boost tabular data classification performance.
problem Improving tabular data classification accuracy.
method Fine-tuning ICL-transformers on synthetic datasets with complex decision boundaries.
result Fine-tuned ICL-transformers outperform regular neural networks on real-world datasets.
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.
This paper proves the theoretical advantage of unsupervised pretraining for machine learning tasks.
problem Understanding why unsupervised pretraining helps in machine learning tasks.
method A generic framework using Maximum Likelihood Estimation (MLE) for unsupervised pretraining and Empirical Risk Minimization (ERM) for downstream tasks.
result Proves an excess risk of i l d e O ( C Φ / m + C Ψ / n ) ilde{\mathcal{O}}(\sqrt{\mathcal{C}_Φ/m} + \sqrt{\mathcal{C}_Ψ/n}) i l d e O ( C Φ / m + C Ψ / n ) for downstream tasks under mild conditions. Data-efficient PDE operator learning without expensive simulations.
problem Expensive numerical PDE solutions limit data efficiency in machine learning.
method Unsupervised pretraining and in-context learning.
result Highly data-efficient and more generalizable than conventional models.
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.
Self-supervised pretraining for heterogeneous hypergraphs improves graph-based tasks.
problem Capturing higher-order relations in heterogeneous hypergraphs.
method SPHH framework for self-supervised pretraining of heterogeneous HyperGNNs.
result SPHH consistently outperforms state-of-the-art baselines in various downstream tasks.
Research improves federated text models for next word prediction.
problem Training models on distributed devices efficiently and effectively.
method Employ transfer learning in federated learning for next word prediction.
result Enhancements to current baselines with pretrained embeddings and whole model pretraining.
In-Run Data Shapley offers efficient data attribution for large-scale models.
problem Existing data attribution methods are computationally intensive and cannot target specific models.
method In-Run Data Shapley, which efficiently attributes data contributions to a specific model without re-training.
result In-Run Data Shapley achieves significant efficiency, enabling data attribution for pretraining models.
The ever-growing volume of data of user-generated content on social media provides a nearly unlimited corpus of unlabeled data even in languages where resources are scarce. In this paper, we demonstrate that state-of-the-art results on two Thai social text categorization tasks can be realized by pretraining a language …
Optimal feature transfer identified through bias-variance analysis.
problem Optimizing feature transfer in transfer learning.
method Simple linear model with fine-grained bias-variance decomposition.
result Optimal pretrained feature transform is naturally sparse.
Multi-role dialogue understanding comprises a wide range of diverse tasks such as question answering, act classification, dialogue summarization etc. While dialogue corpora are abundantly available, labeled data, for specific learning tasks, can be highly scarce and expensive. In this work, we investigate dialogue cont…
This paper evaluates conformal prediction for aerial image classification in challenging environments.
problem Challenging aerial image classification in data-scarce, unconstrained environments.
method Conformal prediction applied to pretrained models (MobileNet, DenseNet, ResNet) with limited labeled data.
result Conformal prediction can provide valuable uncertainty estimates even with small labeled samples.
Transformer pretraining yields strong EB performance without explicit adaptation.
problem Empirical Bayes problems with unknown test distributions.
method Indirect analysis of pretrained transformer's performance under universal priors.
result Near-optimal regret bound of O ~ ( 1 n ) \widetilde{O}(\frac{1}{n}) O ( n 1 ) for arbitrary test distributions. TQA improves prediction intervals for time series data by adjusting quantiles for both cross-sectional and longitudinal coverage.
problem Constructing reliable prediction intervals for cross-sectional time series data.
method Temporal Quantile Adjustment (TQA) method that adjusts the quantile in Conformal Prediction to account for both cross-sectional and longitudinal coverage.
result TQA improves longitudinal coverage while preserving cross-sectional coverage, as validated through extensive experimentation.
ST-BCP narrows the coverage gap in BCP by transforming nonconformity scores.
problem The looseness in BCP's coverage guarantee due to Markov's inequality.
method Introduces a data-dependent transformation of nonconformity scores.
result Reduces the average coverage gap from 4.20% to 1.12% on benchmarks.
Adapts conformal prediction for missing data, ensuring valid coverage.
problem Uncertainty quantification with missing covariates.
method Proposes a reweighted conformal prediction procedure for handling missing values.
result Guaranteed Marginal Coverage and Mask-Conditional Validity for general missing data mechanisms.
LPCI provides valid prediction intervals for longitudinal data.
problem Current conformal prediction methods for time series data lack cross-sectional coverage when applied to longitudinal datasets.
method Modeling residual data as a quantile fixed-effects regression problem, constructing prediction intervals with a trained quantile regressor.
result LPCI achieves valid cross-sectional coverage and outperforms existing benchmarks in terms of longitudinal coverage rates.
We introduce a pretraining technique called Selfie, which stands for SELFie supervised Image Embedding. Selfie generalizes the concept of masked language modeling of BERT (Devlin et al., 2019) to continuous data, such as images, by making use of the Contrastive Predictive Coding loss (Oord et al., 2018). Given masked-o…
TSFMs improve financial forecasting across diverse tasks with strong transferability.
problem Complex nonlinear relationships, temporal dependencies, and limited data in financial time series forecasting.
method Pretraining on diverse time series corpora followed by task-specific adaptation.
result Tiny Time Mixers (TTM) achieved 25-50% better performance on limited data and 15-30% improvements on longer datasets.
Study online conformal prediction for non-stationary data with optimal training-conditional regret.
problem Online prediction for non-stationary data streams with unknown distribution drift.
method Proposes split-conformal and full-conformal algorithms that adapt to drift detection and incorporate stability for online learning.
result Proves minimax-optimal regret for online full conformal algorithm under appropriate restrictions.
Metric learning enhances combinatorial coverage metrics' ability to predict classification errors.
problem Dataset dependence of combinatorial coverage metrics in anticipating classification errors.
method Metric learning to improve latent space separation of data classes.
result Metric learning increases SDCCMs' ability to distinguish between correctly and incorrectly classified data.
Analysis of pretrained models' effectiveness in downstream tasks.
problem Understanding why pretrained models perform well in NLP tasks.
method Analyzed head and prompt tuning approaches using latent variable models.
result Prompt tuning provides stronger guarantees than head tuning.
Framework uses synthetic data from pretrained models to improve predictive modeling.
problem Limited effectiveness of synthetic data from generative models for improving predictive performance.
method Proposes an end-to-end framework that generates and filters synthetic data through domain-specific statistical methods.
result Consistent improvements in predictive performance across various settings.
Transformer models can solve complex math problems with less data.
problem Solving complex symbolic mathematics problems with limited data.
method Pretrain transformer models on language translation tasks and fine-tune for symbolic math.
result Pretrained transformer models achieve comparable accuracy to state-of-the-art models with less data.
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.
Quantile regression undercovers true uncertainty, revealing a bias in high dimensions.
problem Under-coverage bias in uncertainty estimation by quantile regression.
method Theoretical study on coverage of uncertainty estimation algorithms in learning quantiles.
result Quantile regression undercovers true uncertainty, revealing a bias in high dimensions.