RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentation.

[Preprint] Learning to Filter Context for Retrieval-Augmented Generaton

思源码消费年度总结

Processed / Cleaned Data for Paper Copilot

[IJCAI 2024] Papers about graph reduction including graph coarsening, graph condensation, graph sparsification, graph summarization, etc.

Source code for the paper "CAT: Interpretable Concept-based Taylor Additive Models".

Discover and Cure: Concept-aware Mitigation of Spurious Correlation (ICML 2023)

逻辑回归和单层softmax的解析解

Meaningfully debugging model mistakes with conceptual counterfactual explanations. ICML 2022

AI-powered Autonomous Data System

Towards Efficient Shapley Value Estimation via Cross-contribution Maximization

Code and Appendix for Shapley Approximation Methods SVARM and Stratified SVARM (AAAI 2024)

Reduce end to end training time from days to hours (or hours to minutes), and energy requirements/costs by an order of magnitude using coresets and data selection.

Data-efficient Training of Machine Learning Models

Papers for LLM and foundation models for time series analytics

💱 A curated list of data valuation (DV) to design your next data marketplace

Data Shapley: Equitable Valuation of Data for Machine Learning

A fast, effective data attribution method for neural networks in PyTorch

[VLDB 2024] P-Shapley: Shapley Values on Probabilistic Classifiers

Python codes for influential instance estimation

A scalable graph learning toolkit for extremely large graph datasets. (WWW'22, 🏆 Best Student Paper Award)

Graph Neural Scaling Laws