T-Search: An Open Agentic Retriever and Playground for Hard Multi-Step Search
Olga Tsymboi, Ramil Latypov, Aleksandr Medvedev, Danil Taranets, Dmitrii Stoianov
実装難易度
Hard
推論・学習コスト
High
想定用途
生成
概要
We present T-Search, an open-weight agentic retriever for hard multi-step search. Given a question and a search tool over a fixed corpus, it runs a bounded multi-round search and returns a ranked list of evidence chunks with short justifications, leaving answer generation to a downstream model, so backend and generator can be swapped without retraining. T-Search is built on Qwen3.6-35B-A3B and…
何が新しいか
We present T-Search, an open-weight agentic retriever for hard multi-step search. Given a question and a search tool over a fixed corpus, it runs a bounded multi-round search and returns a ranked…
何に使えるか
生成
実装情報
- Paper URL
- あり
実装チェックリスト
実装または配布ページ
要確認Paper onlyの可能性があるため再実装前提で確認してください。
一次情報リンク
OKPaper
検証しやすさ
要確認公式実装が見つからないため、論文から再実装する前提です。
計算資源
要確認学習や高解像度推論ではGPUメモリと実行時間に注意が必要です。
ライセンス
未取得配布元のLICENSE、モデルカード、Paperの利用条件を確認してください。
商用利用
未取得研究利用限定、データセット由来制限、API規約の有無を確認してください。
自社データで試すなら
製造業・材料開発のExcel/CSVデータに落とし込むための最初の手順です。
- 1まず自社データを、入力条件、目的変数、評価したい指標に分けて整理します。
- 2LightGBMやRandom Forestなどのベースラインを先に作り、この手法と比較します。
- 3評価指標はR2/RMSE、AUC、異常検知の再現率、実験回数削減率など、現場の意思決定に近いものを選びます。
- 4SHAPや特徴量重要度で、効いている因子が物理・化学・工程知識と矛盾しないか確認します。
実装難易度
Hard - 公式実装が見つからないため、論文から再実装する前提です。
必要リソース
- GPU目安: High
- データセット: 論文・リポジトリ側の指定を確認してください。
- 学習要否: 再学習や評価環境の準備が必要になる可能性があります。
- 学習や高解像度推論ではGPUメモリと実行時間に注意が必要です。
実務で使う場合の注意点
- ライセンスと商用利用条件は、Paper / GitHub / Hugging Face の配布元で確認してください。
- 精度、再現性、計算コストはデータセットや評価条件に依存します。
- 個人情報や機密データを扱う場合は、入力データの保存先と外部API利用条件を確認してください。
関連記事
Imagine to Act: High-Fidelity Data Synthesis via Image Editing World Model for Scalable GUI Agent Training
Graphical User Interface (GUI) agents have emerged as a promising paradigm for automating complex digital work
Retrieval Is Not Enough: Refreshing Memory for Frozen Time-Series Forecasters
Retrieval-augmented time-series forecasting uses the continuations of historical segments similar to the curre
Learning to Retrieve via Reinforcement Learning in Embedding Space
Dense retrieval models are typically trained with contrastive objectives that learn effective representations
RBMatch: Dual-Level Class Rebalancing for Semi-Supervised Building Footprint Extraction
Accurate building footprint extraction from high-resolution remote sensing imagery is essential for urban plan