A curated list of trustworthy deep learning papers. Daily updating...

A unifying robust reinforcement learning benchmark.

Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities. Arxiv, 2024.

A Systematic Investigation of Transferability and Robustness of Humor Detection Models

Benchmark your model on out-of-distribution datasets with carefully collected human comparison data (NeurIPS 2021 Oral)

PyBullet CartPole and Quadrotor environments—with CasADi symbolic a priori dynamics—for learning-based control and RL

The largest and most challenging benchmark for machine-generated text detectors. (ACL 2024)

Tunable sorting for responsive robustness and beyond!

Out-of-distribution detection, robustness, and generalization resources. The repository contains a professionally curated list of papers, tutorials, books, videos, articles and open-source libraries etc

A toolbox for benchmarking trustworthiness of multimodal large language models (MultiTrust)

A Graph Neural Network-based Intrusion Detection System

Benchmarking Generalized Out-of-Distribution Detection

Find the optimum number of states to use in a ChromHMM model

Specification Curve is a Python package that performs specification curve analysis: exploring how a coefficient varies under multiple different specifications of a statistical model.

DPLL(T)-based Verification tool for DNNs

A Enterprise-Level API Framework based on NestJS, built for scaleability, best-practise and robustness. Based on DDD, Onion, Clean and Hexagonal Architecture

Service to examine data processing pipelines (e.g., machine learning or deep learning pipelines) for uncertainty consistency (calibration), fairness, and other safety-relevant aspects.

A unified evaluation framework for large language models

Generalized Gaussian Temporal Difference Error For Uncertainty-aware Reinforcement Learning

The official implementation of ECCV'24 paper "To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now". This work introduces one fast and effective attack method to evaluate the harmful-content generation ability of safety-driven unlearned diffusion models.