Our Research

Factored enhance CTR predictions through optimized user embeddings and text processing, ranking globally in RecSys.

Factored and Carnegie Mellon tackle dataset bias with constrained statistical inference, ensuring fairness in AI-driven decisions.

Factored and MLCommons deliver the world’s largest multilingual keyword spotting dataset, enhancing speech recognition in 50+ languages.

Factored, MLCommons, and Kaggle enhance generative AI safety by uncovering vulnerabilities through the Adversarial Nibbler Challenge.

Factored and MLCommons deliver scalable, ethically sourced audio for robust speech recognition, transforming AI accessibility globally.

Factored and MLCommons revolutionize medical AI through MedPerf’s scalable federated learning, boosting collaboration across 50+ global institutions.

Factored, MLCommons, and NVIDIA tackle dataset redundancy with scalable LSH models, presented at NeurIPS 2021.

Factored and MLCommons transform data-centric AI through DataPerf’s benchmarks, driving innovations in dataset quality and evaluation.

Factored, NVIDIA, and MLCommons drive multilingual AI innovation with 77-language speech datasets, promoting ethical data and global inclusivity.

Factored and MLCommons scale adversarial testing for T2I models, ensuring safe, ethical generative AI outputs.

Factored, MLCommons, and Hugging Face drive AI inclusivity through multicultural feedback, aligning LLMs with human values.