AI Paper Atlas

Research map / Large language models / Agents, code and math

ChinaTravel: An Open-Ended Travel Planning Benchmark with Compositional Constraint Validation for Language Agents

Jie-Jing Shao, Bo-Wen Zhang, Xiao-Wen Yang, Baizhi Chen, Siyu Han, Pang Jinghao, Wen-Da Wei, Guohao Cai, Zhenhua Dong, Lan-Zhe Guo, Yu-Feng Li

ICLR 2026 · Large language models › Agents, code and math · approach: Long-horizon LLM agents · 32 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use ground-truth (oracle) DSL instead of LLM NL2DSL translation (ablation)NeSy Planning (DeepSeek-V3) · ChinaTravel EasyFinal Pass Rate: 52.6 -> 74.0Improved
use ground-truth DSL instead of LLM translation (ablation)NeSy Planning (DeepSeek-V3) · ChinaTravel Human-ValFinal Pass Rate: 37.0 -> 45.4Improved
use ground-truth DSL instead of LLM translation (ablation)NeSy Planning (DeepSeek-V3) · ChinaTravel Human-TestFinal Pass Rate: 23.3 -> 31.6Improved
replace pure-LLM ReAct with neuro-symbolic planningDeepSeek-V3 · ChinaTravel EasyFinal Pass Rate: 5.33 -> 52.6Improved
replace ReAct with NeSy PlanningGPT-4o · ChinaTravel EasyFinal Pass Rate: 0 -> 60.6Improved
use NeSy Planning instead of TTG MILP solverDeepSeek-V3 · ChinaTravel EasyFinal Pass Rate: 8.66 -> 52.6Improved
add LLM-modulo oracle verifier feedback loop (10 rounds)LLM-based planning · ChinaTravel EasyFinal Pass Rate: 4.33 (DeepSeek), 7.00 (GPT-4o)No clear effect
add 1-shot example to ReActDeepSeek-V3 · ChinaTravel Human-ValFinal Pass Rate: 0 -> 2.59No clear effect
add syntax-checking Reflexion to NL2DSLLLMs in NeSy Planning · DSL translationconstraint countHurt
use LLM-enhanced preference recommendation (PEQ) over baseline (BQ) (ablation)NeSy Planning · single-preference subsetspreference optimization: better in 5/6 scenariosMixed
use PEQ over BQ for two preferences (ablation)NeSy Planning · multi-preference trade-offsPairwise Pareto win rate: 0.58 (P3&P4), 0.45 (P1&P2), 0.10 (P1&P4)Mixed
move from literal to semantic POIs in intent groundingDeepSeek-V3 / GPT-4o · Intent Groundingaccuracy: DS 94% -> 76%, GPT 79% -> 53%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy