What’s Next in AI is Fluid Intelligence
What’s Next in AI is Fluid Intelligence
Today's AI is narrow. Applying trained models to new challenges requires an immense amount of new data training, and time.
We need AI that combines different forms of knowledge, unpacks causal relationships, and learns new things on its own.
In short, AI must have fluid intelligence— and that's exactly what AI research teams are building.
Today's AI is narrow. Applying trained models to new challenges requires an immense amount of new data training, and time.
We need AI that combines different forms of knowledge, unpacks causal relationships, and learns new things on its own.
In short, AI must have fluid intelligence— and that's exactly what AI research teams are building.
Workstreams
Neurosymbolic AI
We're integrating neural and symbolic techniques to build AI that can perform complex tasks by understanding and reasoning more like we do.
AI Hardware
Our digital and analog accelerators are driving massive improvements in computational power while remaining energy-efficient.
Secure, Trusted AI
Trust and security should be baked into the core of any AI we put out into the world. We're building tools to help you ensure that it is.
AI Engineering
We're building tools to help AI creators reduce the time they spend training, maintaining, and updating their models.
Publication collections
|
Dec 2020 |
Conference on Neural Information Processing Systems (NeurIPS) |
|
|
Aug 2020 |
ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) |
|
|
Jul 2020 |
Association for Computational Linguistics (ACL) |
|
|
Feb 2020 |
Association for the Advancement of Artificial Intelligence (AAAI) |
|
|
Dec 2019 |
Conference on Neural Information Processing Systems (NeurIPS) |
Featured
Journey inside a new class of Analog AI hardware
At IBM Research we’re developing a new class of Analog AI hardware, purpose built to help innovators realize the promise of the next stages of AI.
Recent news
Blog
Getting AI to Reason: Using Neuro-Symbolic AI for Knowledge-Based Question Answering
4-Dec-2020
Blog
IBM Teams with Industry Partners to Bring Energy-Efficient AI Hardware to Hybrid Cloud Environments
21-Oct-2020
Press Release
How to build trust in your AI model: the latest innovation from IBM Watson
9-Dec-2020
Experiments
The Open Source python toolkit for exploring and using the capabilities of in-memory computing devices in the context of artificial intelligence.
Compare VSRL with traditional reinforcement learning to see how they perform under different environmental conditions and with different amounts of training.
Try CLAI, an open-source framework for AI-powered command line plugins. CLAI helps you navigate the command line more efficiently, removing roadblocks and finding missing dependencies.
Publications
| Date | Content | Title | Journal / Venue |
|---|---|---|---|
| Jun 2020 | Paper |
Verifiably Safe Exploration for End-to-End Reinforcement Learning |
arXiv |
| May 2020 | Paper |
Accurate deep neural network inference using computational phase-change memory |
Nature Communications |
| Jan 2020 | Paper |
Towards a Homomorphic Machine Learning Big Data Pipeline for the Financial Services Sector |
IACR |
AI research teams
| AI Hardware | |
| Algorithmic Acceleration | |
| Auto AI (tools) | |
| Computer Vision | |
| Explainability | |
| Fairness | |
| Knowledge and Reasoning |
| Machine Learning | |
| Natural Language | |
| Process Automation | |
| Robustness | |
| Speech | |
| Transparency and Accountability | |
| Value Alignment |
