Mike Murphy

Kim Martineau

News

Research

Explainer

Saška Mojsilović wants to channel AI for good. She may also make you rethink sour cabbage

What is prompt-tuning?

IBM’s Stacy Hobson wants to build tech that works for everyone

What is AI alignment?

The latest AI safety method is a throwback to our maritime past

What is red teaming for generative AI?

IBM reaffirms its commitment to the Rome Call for AI ethics

We're developing technological solutions to assist subject matter experts with their scientific workflows by enabling the Human-AI co-creation process.  

Accelerator Technologies

IJCAI 2024

ACL 2024

MICCAI 2024

NeurIPS 2024

Trustworthy AI

Fairness, Accountability, Transparency

Adversarial Robustness and Privacy

AI Transparency

Responsible Technology

Foundation Models

Automated AI

Computer Vision

Human-Centered AI

Machine Learning

Natural Language Processing

Generative AI

Security

Data and AI Security

Granite

Explainable AI

AI Testing

Causal modeling is crucial to the effectiveness and trust of AI by ensuring that actions lead to intended outcomes. We study the inference of causal effects and relationships, as well as the application of causal thinking to out-of-distribution generalization, fairness, robustness, and explainability.

Causality

Even advanced AI systems can be vulnerable to adversarial attacks. We’re making tools to protect AI and certify its robustness, including quantifying the vulnerability of neural networks and designing new attacks to make better defenses. And we’re helping AI systems adhere to privacy requirements.

We’re making tools to protect AI and certify its robustness, and helping AI systems adhere to privacy requirements.

The combination of increasingly powerful computers and AI offers the possibility to be able to detect, diagnose, and cure diseases like never before. At IBM Research, we’re working on creating software and AI systems that can convert reams of health data into useable information for clinicians the world over.

Healthcare and Life Sciences

We're building systems that can convert reams of health data into useable information for clinicians the world over.

To trust AI systems, explanations can go a long way. We’re creating tools to help debug AI, where systems can explain what they’re doing. This includes training highly optimized, directly interpretable models, as well as explanations of black-box models and visualizations of neural network information flows.

We’re creating tools to help AI systems explain why they made the decisions they did. 

Biases can lead to systematic disadvantages for marginalized individuals and groups — and they can arise in any point in the AI development lifecycle. To increase the accountability of high-risk AI systems, we're developing technologies to increase their end-to-end transparency and fairness.

Saška Mojsilović at IBM Research in Yorktown

Fairness, Accountability, Transparency

Our work

IBM reaffirms its commitment to the Rome Call for AI ethics

What is red teaming for generative AI?

The latest AI safety method is a throwback to our maritime past

What is AI alignment?

IBM’s Stacy Hobson wants to build tech that works for everyone

What is prompt-tuning?

Projects

Accelerator Technologies

Publications

Better Bias Benchmarking of Language Models via Multi-factor Analysis

Interpolating Item and User Fairness in Multi-Sided Recommendations

Protecting Users From Themselves: Safeguarding Contextual Privacy in Interactions with Conversational Agents

Impact of Skin Tone Diversity on Out-of-Distribution Detection Methods in Dermatology

Biasly: An Expert-Annotated Dataset for Subtle Misogyny Detection and Mitigation

Using Causal Inference to Investigate Contraceptive Discontinuation in Sub-Saharan Africa

Related topics