Instruction

Large language models (LLMs) have achieved remarkable success, yet they face significant challenges in logical and symbolic reasoning. This is because (i) learning syntax, semantics, and world knowledge through tasks such as next-word prediction or masked language modeling does not ensure the logical reasoning ability of LLMs, and (ii) the pre-training corpora of LLMs primarily consist of human-written texts, which lack high-quality logical reasoning samples such as logical deduction and proofs. This two-day bridge program aims to thoroughly explore and expand the intersection of AI and Logic, with special interest in logical and symbolic reasoning in language models. This would improve the reasoning capabilities of LLMs to solve complex logical problems requiring sophisticated deductive, inductive, or abductive reasoning, and to avoid producing responses that contradict themselves across multiple relevant questions.

Topics

Topics of interest include, but are not limited to:

- Logical question answering of LLMs;
- Chain-of-thought reasoning of LLMs;
- External tool-use (e.g., logic solvers) for LLM reasoning;
- Logical consistency of LLMs (e.g., implication, negation);
- Symbolic expressions and reasoning of LLMs;
- Multi-turn interactions for LLM reasoning;
- Multi-agent LLMs’ reasoning;
- Benchmarks and evaluation for logical reasoning of LLMs;
- Relevant topics from logic, linguistics, and cognitive science.

Important Dates

First-Round Submission Deadline: October 31, 2025
First-Round Notification of Acceptance: November 14, 2025
Second-Round Submission Deadline: December 5, 2025
Second-Round Notification of Acceptance: December 20, 2025

All deadlines are specified in Anywhere on Earth (AoE).

Submission Guidelines

We welcome two types of papers:

Full papers: Full-length research papers from 4 to 8 pages (excluding references and appendices);
Short papers: research/position papers of up to 4 pages (excluding references and appendices).

Papers should be submitted in the AAAI format (https://aaai.org/authorkit26-1/). The review process will be single-blinded, and we welcome accepted and published papers. The contributions can be either non-archival or published upon the authors' choices. There will be Oral Presentation Awards, Outstanding Paper Awards, and one Best Paper Award for accepted papers with outstanding quality.

Submission Site

The bridge program uses OpenReview for paper submission and reviewing. The submission link is https://openreview.net/group?id=AAAI.org/2026/Bridge/LMReasoning.

Attendence

The bridge program is open to researchers, practitioners, and industry professionals interested in logical and symbolic reasoning of language models. There are no specific criteria or maximum number of attendees.

Format of the Bridge

This is a two-day bridge program with no restrictions on attendance. The bridge program is organized to include keynote presentations of invited speakers, panel discussion, tutorials, oral paper presentations, and poster sessions.