Neuro-Symbolic AI Summer School 2025

​Tired of the same old AI ideas and looking for something new? Or wondering if there are any deep alternatives to models that hallucinate, elude human debuggability, and have only limited ability to reason, plan, or do math?

​📢 Please join us for the 4th Neuro-Symbolic AI Summer School (NSSS), a free, fully-remote meeting to be held Aug 14-15, 2025 - where you can learn how to be part of this exciting area at the most forward-looking cutting edge of AI. This year's theme is "AI for Precise Computation: Mathematics, Reasoning, and Planning".

​🌐 NSSS ‘25 is the latest in a series of research and educational events that aim to accelerate progress in the fast-emerging area of neuro-symbolic AI by teaching graduate students, data scientists, and researchers principles from the side of AI they may be less familiar with, as well as presenting a curation of emerging research ideas at the intersection.

​🎯 The focus of the event is on techniques, in particular those which augment neural network/ML ideas with symbolic AI ideas, to address at least three main open problems of AI: 1) human interpretability/controllability, 2) learning with less data/computation (e.g. via knowledge), and 3) out-of-distribution generalization (e.g. via reasoning).

📅 Agenda

​Day 1: Frameworks and Foundations | AUG 14

📜 Big Picture
9:00–9:30 AM ET – Alexander Gray, Centaur AI Institute – Neuro-Symbolic AI and NSSS4 (Overview)
9:30–9:40 AM – Q&A / Break
9:40–10:20 AM – Arvind Narayanan, Princeton – Assessing the Current State of AI (Talk)
10:20–10:30 AM – Q&A / Break

​🏗 Frameworks
10:30–11:05 AM – Leilani Gilpin, UC Santa Cruz – Neuro-Symbolic AI for Safer Autonomous Vehicles (Tutorial)
11:05–11:15 AM – Q&A / Break
11:15–12:00 PM – Mateo Zarlenga & Pietro Barbiero, Cambridge – Foundations of Interpretable Models (Tutorial)
12:00–12:10 PM – Q&A / Break
12:10–12:55 PM – Parisa Kordjamshidi, Michigan State – Compositional Learning in Language and Vision (Tutorial)
12:55–1:05 PM – Q&A / Break

​📐 Mathematical Foundations
1:05–1:45 PM – Parikshit Ram, IBM Research – How to Measure Compositionality, and Why it Leads to Better Generalization (Talk)
1:45–1:55 PM – Q&A / Break
1:55–2:35 PM – Peihao Wang, UT Austin – Why Neural Networks Can Discover Symbolic Structures (Talk)
2:35–2:45 PM – Q&A / Break
2:45–3:25 PM – Changlong Wu, Purdue – Why Current Models Will Always Hallucinate (and a Path Forward) (Talk)
3:25–3:35 PM – Q&A / Break
3:35–4:35 PM – Sridhar Mahadevan, Adobe Research & UMass Amherst – Category Theory: The Mathematics of Symbolic Structures (Tutorial)
4:35–4:45 PM – Q&A / Break

​Day 2: Methods and Systems | AUG 15

​🤖 Neuro-Symbolic AI Software
9:00–9:40 AM – Bowen Li, Carnegie Mellon – Generalizing to New Situations in Robotics (Talk)
9:40–9:50 AM – Q&A / Break
9:50–10:30 AM – Steve Carrow, IBM; Olga Vilenskaia, Centaur AI Institute;  Kevin O'Connor, Centaur AI Institute – Explainability While Retaining Predictive Accuracy (Talk)
10:30–10:40 AM – Q&A / Break

​🧩 Learning Symbolic Models
10:40–11:20 AM – Bin Yu, UC Berkeley – Interpretable and Veridical Data Science (Talk)
11:20–11:30 AM – Q&A / Break
11:30–12:10 PM – Felix Petersen, Stanford – Deep Differentiable Logic Gate Networks (Talk)
12:10–12:20 PM – Q&A / Break
12:20–1:00 PM – Hikaru Shindo, TU Darmstadt – Neuro-symbolic Agentic Systems (Talk)
1:00–1:10 PM – Q&A / Break

​🛡 Safer AI Systems
1:10–1:50 PM – Soroush Saghafian, Harvard – Human + AI "Centaur" Systems (Talk)
1:50–2:00 PM – Q&A / Break
2:00–2:40 PM – Pranava Madhyastha, City Univ London – New Results in Controllable Text Generation (Talk)
2:40–2:50 PM – Q&A / Break

​➗ AI Systems for Mathematics
2:50–3:30 PM – Shange Tang, Princeton – State-of-the-art Performance in Automated Mathematical Theorem Proving (Talk)
3:30–3:40 PM – Q&A / Break
3:40–4:20 PM – Ankit Anand, DeepMind – Curious Case of AI in Maths: Being Proficient in Advancing Open Conjectures in Maths Yet Having Struggles in AI for Education (Talk)
4:20–4:30 PM – Q&A / Break

​🔮 Looking Forward
4:30–5:30 PM – Panel on The Future of AI – Rich Sutton, Univ Alberta; Leonardo de Moura, Amazon; Artur Garcez, City Univ London; more TBA; moderator: Alexander Gray – Discussion including open Q&A
5:30–5:35 PM – Alexander Gray, Centaur AI Institute – What's Coming Next and How to Participate (Closing remarks)

​………………………………….

​Background on previous events:

​Our previous Neuro-Symbolic AI Summer School events have attracted thousands of registrants and featured top researchers including Turing Award winners and National Academies members.