Advancing AI Systems
Robust, Trustwory Solutions

As leading AI researchers, Miniml push the boundaries of Machine Learning, NLP, and Generative AI. Our work advances the state of AI, focusing on mitigating hallucinations, enhancing model accuracy, and ensuring data security across diverse industries.

Using Natural Language Explanations
to Improve Robustness of In-context Learning

AI models are improving in decision-making, but their reliability against adversarial inputs remains a challenge. Miniml’s research shows how integrating human-like natural language explanations can strengthen model robustness, delivering more accurate results across complex tasks.

Advancing AI Systems
Robust, Trustwory Solutions

Using Natural Language Explanations
to Improve Robustness of In-context Learning

Selected Publications

GRADA: Graph-based Reranker against Adversarial Documents Attack

FLARE: Faithful Logic-Aided Reasoning and Exploration

DeCoRe: Decoding by Contrasting Retrieval Heads to Mitigate Hallucinations

Neurosymbolic Diffusion Models

MMLongBench: Benchmarking Long-Context Vision-Language Models Effectively and Thoroughly

Advancing AI Systems Robust, Trustwory Solutions

Using Natural Language Explanations to Improve Robustness of In-context Learning

Selected Publications

GRADA: Graph-based Reranker against Adversarial Documents Attack

FLARE: Faithful Logic-Aided Reasoning and Exploration

DeCoRe: Decoding by Contrasting Retrieval Heads to Mitigate Hallucinations

Neurosymbolic Diffusion Models

MMLongBench: Benchmarking Long-Context Vision-Language Models Effectively and Thoroughly

Advancing AI Systems
Robust, Trustwory Solutions

Using Natural Language Explanations
to Improve Robustness of In-context Learning