Pedro Domingos: Tensor Logic Unifies AI Paradigms

168 snips

Dec 8, 2025

Pedro Domingos, a leading computer science professor at the University of Washington and author of The Master Algorithm, unveils his groundbreaking concept, TensorLogic. He discusses how this innovative programming language could unify the fragmented worlds of Deep Learning and Symbolic AI. Pedro reveals TensorLogic's capabilities in logical reasoning and learning from data, emphasizing its potential to prevent AI hallucinations. He also shares insights on how TensorLogic can express complex systems and improve AI education, paving the way for a more integrated future in artificial intelligence.

Ask episode

AI Snips

Chapters

Books

Transcript

Episode notes

ADVICE

Optimize Einsum With A CUDA Back End

Implement the tensor-equation abstraction efficiently on CUDA to make einsum-style programs practical and fast.
Optimize a single high-level construct and you unlock performance and developer adoption over clunky einsum syntax.

INSIGHT

Learning New Predicates Via Tensor Decomposition

Structural learning and predicate invention fall out from gradient descent when rules are written as tensor decompositions.
Training discovers compact latent predicates analogous to object concepts by tensor factorization and thresholding.

INSIGHT

Use Soft Equations As Architectural Priors

TensorLogic supports broad, soft priors as equations so gradient descent can learn architectures within a constrained space.
Developers iteratively refine equations instead of fully hard‑coding architectures, improving interpretability and search.

Get the Snipd Podcast app to discover more snips from this episode

Get the app