Responsible Multimodal AI
Counterfactual datasets and evaluation methods for measuring social, cultural, and intersectional biases in vision-language models.
AI research leadership · Thoughtworks AI Labs
Phillip Howard
I lead a team of AI researchers making generative AI more reliable through pathfinding research on evals, interpretability, and model robustness. My recent work focuses on multimodal models, synthetic data generation, counterfactual evaluation, temporal reasoning, responsible AI, and model interpretability. Before joining Thoughtworks, I spent nearly a decade at Intel working on multimodal cognitive AI, vision-language models, natural language processing, and applied machine learning.
Latest
News
- Tutorial on Counterfactual Fairness Analysis in Language-Vision Models accepted to ACM FAccT 2026.
- Three papers were accepted to ICLR 2026: p-less Sampling (oral), Scaling Knowledge Graph Construction through Synthetic Data Generation and Distillation, and Is Your Paper Being Reviewed by an LLM?
- Learning from Reasoning Failures via Synthetic Data Generation was accepted to AAAI 2026.
- Two papers were accepted to EMNLP 2025: Pruning the Paradox and Transformer-Based Temporal Information Extraction and Application: A Review.
- A Semantic Parsing Framework for End-to-End Time Normalization was accepted to NeurIPS 2025.
- SK-VQA was presented at ICML 2025 (oral).
- Two papers were accepted to NAACL 2025: Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals (oral) and LVLM-Compress-Bench.
Focus Areas
Synthetic Data
Targeted data generation for reasoning, retrieval-augmented multimodal generation, and robustness under distribution shift.
Temporal & Structured Reasoning
Models and representations for temporal question answering, time normalization, knowledge graphs, and structured reasoning.
Interpretability
Tools and methods for interpreting CLIP-like models, sparse autoencoders for vision models, and nonlinear representation learning.
Selected Papers
Cross-Cultural Value Awareness in Large Vision-Language Models
Phillip Howard, Xin Su, Kathleen C. Fraser
Cultural Counterfactuals: Evaluating Cultural Biases in Large Vision-Language Models with Counterfactual Examples
Phillip Howard, Xin Su, Kathleen C. Fraser
p-less Sampling: A Robust Hyperparameter-Free Approach for LLM Decoding
Runyan Tan, Shuang Wu, Phillip Howard
Scaling Knowledge Graph Construction through Synthetic Data Generation and Distillation
Prafulla Kumar Choubey, Xin Su, Man Luo, Xiangyu Peng, Caiming Xiong, Tiep Le, Shachar Rosenman, Vasudev Lal, Phil L. Mui, Ricky Ho, Phillip Howard, Chien-Sheng Wu
Is Your Paper Being Reviewed by an LLM? Benchmarking AI Text Detection in Peer Review
Sungduk Yu, Man Luo, Avinash Madasu, Vasudev Lal, Phillip Howard
Learning from Reasoning Failures via Synthetic Data Generation
Gabriela Ben Melech Stan, Estelle Aflalo, Avinash Madasu, Vasudev Lal, Phillip Howard
Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals
Phillip Howard, Kathleen C. Fraser, Anahita Bhiwandiwalla, Svetlana Kiritchenko
LVLM-Compress-Bench: Benchmarking the Broader Impact of Large Vision-Language Model Compression
Souvik Kundu, Anahita Bhiwandiwalla, Sungduk Yu, Phillip Howard, Tiep Le, Sharath Nittur Sridhar, David Cobbley, Hao Kang, Vasudev Lal
Pruning the Paradox: How CLIP's Most Informative Heads Enhance Performance While Amplifying Bias
Avinash Madasu, Vasudev Lal, Phillip Howard
Transformer-Based Temporal Information Extraction and Application: A Review
Xin Su, Phillip Howard, Steven Bethard
SK-VQA: Synthetic Knowledge Generation at Scale for Training Context-Augmented Multimodal LLMs
Xin Su, Man Luo, Kris W. Pan, Tien Pei Chou, Vasudev Lal, Phillip Howard
A Semantic Parsing Framework for End-to-End Time Normalization
Xin Su, Sungduk Yu, Phillip Howard, Steven Bethard
SocialCounterfactuals: Probing and Mitigating Intersectional Social Biases in Vision-Language Models with Counterfactual Examples
Phillip Howard, Avinash Madasu, Tiep Le, Gustavo Lujan Moreno, Anahita Bhiwandiwalla, Vasudev Lal
NeuroComparatives: Neuro-Symbolic Distillation of Comparative Knowledge
Phillip Howard, Junlin Wang, Vasudev Lal, Gadi Singer, Yejin Choi, Swabha Swayamdipta
NeuroPrompts: An Adaptive Framework to Optimize Prompts for Text-to-Image Generation
Shachar Rosenman, Vasudev Lal, Phillip Howard
Semi-Structured Chain-of-Thought: Integrating Multiple Sources of Knowledge for Improved Language Model Reasoning
Xin Su, Tiep Le, Steven Bethard, Phillip Howard
COCO-Counterfactuals: Automatically Constructed Counterfactual Examples for Image-Text Pairs
Tiep Le, Vasudev Lal, Phillip Howard
Fusing Temporal Graphs into Transformers for Time-Sensitive Question Answering
Xin Su, Phillip Howard, Nagib Hakim, Steven Bethard
Cross-Domain Aspect Extraction using Transformers Augmented with Knowledge Graphs
Phillip Howard, Arden Ma, Vasudev Lal, Ana Paula Simoes, Daniel Korat, Oren Pereg, Moshe Wasserblat, Gadi Singer
NeuroCounterfactuals: Beyond Minimal-Edit Counterfactuals for Richer Data Augmentation
Phillip Howard, Gadi Singer, Vasudev Lal, Yejin Choi, Swabha Swayamdipta
TempoQR: Temporal Question Reasoning over Knowledge Graphs
Costas Mavromatis, Prasanna Lakkur Subramanyam, Vassilis N. Ioannidis, Adesoji Adeshina, Phillip Howard, Tetiana Grinberg, Nagib Hakim, George Karypis
InterpreT: An Interactive Visualization Tool for Interpreting Transformers
Vasudev Lal, Arden Ma, Estelle Aflalo, Phillip Howard, Ana Simoes, Daniel Korat, Oren Pereg, Gadi Singer, Moshe Wasserblat
Experience
2025-present
AI Researcher, Team Lead & Manager, Thoughtworks
Leading a team of AI researchers focused on applied and frontier AI research.
Leading a team of AI researchers focused on applied and frontier AI research.
2016-2025
Senior Staff AI Research Scientist, Team Lead & Manager, Intel Labs
Led a team of AI researchers focused on multimodal cognitive AI.
Led a team of AI researchers focused on multimodal cognitive AI.
Ph.D.
Arizona State University
Industrial Engineering. Dissertation: Distinct Feature Learning and Nonlinear Variation Pattern Discovery Using Regularized Autoencoders.
Industrial Engineering. Dissertation: Distinct Feature Learning and Nonlinear Variation Pattern Discovery Using Regularized Autoencoders.