TRACE: Trajectory Selection for Parallel Scaling of Search Agents
cs.LG, cs.AI
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/Jaasssoooonnnnn/TRACE
Terminology
Sources
- Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
- BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agent
- Inductive Representation Learning on Large Graphs
- Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
- Fact, Fetch, and Reason: A Unified Evaluation of Retrieval-Augmented Generation
- Agentic Aggregation for Parallel Scaling of Long-Horizon Agentic Tasks
- ParallelMuse: Agentic Parallel Thinking for Deep Information Seeking
- Let's Verify Step by Step
- When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories
- Budget-aware Test-time Scaling via Discriminative Verification
- MuSiQue: Multihop Questions via Single-hop Question Composition
- Mapping the Minds of LLMs: A Graph-Based Analysis of Reasoning LLM
- ReAct: Synergizing Reasoning and Acting in Language Models
- Pushing Test-Time Scaling Limits of Deep Search with Asymmetric Verification
- Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models
- Scaling Test-time Compute for LLM Agents
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks