把智能体的探索过程变成可查询、可复用的数据资产。
Experience Graphs: The Data Foundation for Self-Improving Agents
- 将智能体的搜索过程建模为可查询的体验图，作为核心数据库状态。
- 跨会话复用使优化速度提升10倍，令牌消耗降低52%。
- 适合需要长期积累经验的智能体系统，如代码生成与科研发现。
长周期智能体任务（如代码生成、科学发现、硬件设计）需要持续探索：生成成果、调用工具、观察失败、分支修复，历经数百步。这一过程产生结构化对象——体验图，包含可执行产物、工具输出、奖励、兄弟比较和因果链。然而现有框架将这些体验视为一次性状态，仅以JSON检查点和日志形式存在，无法恢复、跨用户查询或转化为训练数据。我们提出Trellis：一个将体验图作为第一类数据管理的数据库基础架构。核心洞察是：对体验图的搜索即数据库访问模式。前沿选择是查询，跨会话复用是向量种子图检索，训练数据提取是物化视图，还原任一历史步骤的知识则是时间旅行查询。当数据库掌控体验图时，智能体变为无状态计算，崩溃恢复、横向扩展和闭环训练飞轮自然成为架构副产品。我们在Meta的生产级加速器内核优化器KernelEvolve中验证了该设计，跨会话复用实现目标加速比约10倍，令牌成本降低52%。更广泛地，Trellis将推理时搜索从一次性计算转变为持久机构资产：日志使数据库可靠，体验图或使智能体具备累积性。
原文摘要 · Abstract (English)
The database community has repeatedly advanced the state of the art by recognizing that new workloads demand new system architectures. We argue that long-horizon agentic tasks -- code generation, scientific discovery, hardware design -- are such a workload. These agents explore: they generate artifacts, execute tools, observe failures, branch, and repair over hundreds of steps. This search produces a structured object we call an experience graph: executable artifacts, tool outputs, rewards, sibling comparisons, and causal lineage. Yet existing agent frameworks treat this experience as disposable state -- JSON checkpoints and session logs that cannot be recovered after a crash, queried across users, or materialized into training data. We propose Trellis: a data foundation that treats the experience graph as first-class, governed, queryable database state. The core insight is that search over experience graphs is a database access pattern. Frontier selection is a query, cross-session reuse is vector-seeded graph retrieval, training-data extraction is a materialized view, and reconstructing what an agent knew at any past step is a time-travel query. When the database owns the experience graph, agents become stateless compute, and crash recovery, horizontal scaling, and a closed-loop training flywheel emerge as architectural byproducts. We ground the design in KernelEvolve, a production accelerator-kernel optimizer at Meta, where cross-session reuse reaches a target speedup roughly 10x faster at 52% lower token cost. More broadly, Trellis turns inference-time search from disposable computation into a durable institutional asset: logs made databases reliable; experience graphs may make agents cumulative.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。