temporal-reasoning

DreamGraph is a graph-governed conceptual development environment (CDE) that turns plans, architecture decisions, and project knowledge into auditable execution through a persistent cognitive graph.

A Survey of Task-Oriented Knowledge Graph Reasoning: Status, Applications, and Prospects

Time-R1: Framework and resources for endowing LLMs with comprehensive temporal reasoning (understanding, prediction, creative generation) using a novel three-stage RL curriculum. Includes the Time-Bench dataset and pre-trained models.

[WSDM 2022] EvoKG: Jointly Modeling Event Time and Network Structure for Reasoning over Temporal Knowledge Graphs

Qualitative Reasoning: Spatio-Temporal Reasoning using Relation Algebras and Constraint Networks. Documentation is at ReadTheDocs. See link below.

The official repo of TimeLlama, an instruction-finetuned Llama2 series that improve complex temporal reasoning ability.

Reproducible LLM forecasting with a fixed 300-question benchmark, Score and time-masked retrieval.

The repository for ACL 2024 paper "TimeBench: A Comprehensive Evaluation of Temporal Reasoning Abilities in Large Language Models"

[AAAI 2022] TempoQR: Temporal Question Reasoning over Knowledge Graphs

Code and data for "Living in the Moment: Can Large Language Models Grasp Co-Temporal Reasoning?" (ACL 2024)

TRAM: Benchmarking Temporal Reasoning for Large Language Models (Findings of ACL 2024)

Code and data for "Timo: Towards Better Temporal Reasoning for Language Models" (COLM 2024)

Long-term memory for LLM apps that shows its work: dated facts, as-of answers, supersession, and receipts on every answer. Library, CLI, local GUI, MCP.

a contextualized first-order logic knowledge base with a justified-inference reasoning engine

Perfect memory. Exact reasoning. Complete history for AI agents.

[ACL 2025] Learning to Reason Over Time: Timeline Self-Reflection for Improved Temporal Reasoning in Language Models

A temporal visual reasoning benchmark over egocentric robot manipulation video — 39 scenarios, 7 dimensions, 34,713 questions

[EMNLP 2023] Question Answering as Programming for Solving Time-Sensitive Questions

No longer developed. Last release 0.5.3 (MIT). FreshContext services: https://freshcontext.dev

[EMNLP 2026 Findings] Video-MME-Logical: A Controlled Diagnostic Benchmark for Video Temporal-Logical Reasoning