日本フィジカルAI新聞

世界のフィジカルAIを、日本語で。

自動運転/説明可能AIarXiv:2605.16087

信頼性と説明可能性を備えたAI知覚モデルへ:概念からプロトタイプ車両への展開

Towards Trustworthy and Explainable AI for Perception Models: From Concept to Prototype Vehicle Deployment

シェア:XThreadsFacebookLINEはてブBluesky

自動運転の3D知覚モデルに、説明可能性・不確実性推定・ロバスト性を統合し、実車両でリアルタイムに可視化するシステムを構築した論文。

詳しい要約

1. どんなもの?

- 提案: Trustworthy AI perception module for 3D scene understanding in Autonomous Driving. - 特徴: robust, faithful explainability, calibrated uncertainty estimates. - 基盤: transformer-based detector. - 展開: prototype vehicle with XAI Interface for real-time monitoring.

2. 先行研究と比べてどこがすごい?

- 先行: theoretical frameworks for safe and Explainable AI (XAI) exist. - 課題: concrete implementations for 3D scene understanding remain scarce. - 貢献: 実装とprototype vehicle deploymentを実現.

3. 技術・手法の肝は?

- 説明: attention mechanism at inference time. - 検証: perturbation-based consistency tests for faithfulness. - 不確実性: uncertainty estimation and calibration module. - 頑健性: robustness-enhancing training methods.

4. どうやって有効だと検証した?

- 実験: faithful saliency behavior, improved robustness, well-calibrated uncertainty estimates. - 展開: prototype vehicle with XAI Interface visualizing documentation artifacts, model uncertainty state, saliency maps. - 実証: feasibility of trustworthy perception monitoring in real time.

5. 議論はある?

- 言及: opacity conflicts with Trustworthy AI guidelines, complicates safety assurance, debugging, human oversight. - 限界: 要旨からは不明.

6. 次に読むべき論文は?

- 参照: transformer-based detector, Explainable AI (XAI), Trustworthy AI guidelines. - 関連: perturbation-based consistency tests, uncertainty estimation and calibration.

※ AIが要旨から生成した要約です。正確性は原文をご確認ください。

著者: Till Beemelmanns, Shayan Sharifi, Manas Mehrotra, Ayushman Choudhuri, Lutz Eckstein

分類: cs.RO, cs.AI

原文アブストラクト

Deep Neural Networks have become the dominant solution for Autonomous Driving perception, but their opacity conflicts with emerging Trustworthy AI guidelines and complicates safety assurance, debugging, and human oversight. While theoretical frameworks for safe and Explainable AI (XAI) exist, concrete implementations of Trustworthy AI for 3D scene understanding remain scarce. We address this gap by proposing a Trustworthy AI perception module that is remarkably robust, integrates faithful explainability, and calibrated uncertainty estimates. Building on a transformer-based detector, we derive explanation from the attention mechanism at inference time and validate their faithfulness using perturbation-based consistency tests. We further integrate an uncertainty estimation and calibration module, and apply robustness-enhancing training methods. Experiments show faithful saliency behavior, improved robustness, and well-calibrated uncertainty estimates. Finally, we deploy these Trustworthy AI elements in a prototype vehicle and provide an XAI Interface that visualizes documentation artifacts, model uncertainty state, and saliency maps, demonstrating the feasibility of trustworthy perception monitoring in real time. Supplementary materials are available at https://tillbeemelmanns.github.io/trustworthy_ai/ .

関連論文

  1. 1Boston Dynamics、車輪付きAtlasの研究試作機を紹介研究・技術
  2. 2GXOとApptronik、Apolloの物流現場試験を拡大 最大1,000台の展開を構想量産・展開
  3. 3newmoとKDDIが資本業務提携 2028年に堺市で自動運転タクシー商用化目指す提携・買収
  4. 4NVIDIA、CoRLでGR00T 2とIsaac Lab 3.0を発表新製品・発表
  5. 5UC BerkeleyとNVIDIAの研究者、AIエージェントでロボット制御コードを仮想環境で改善研究・技術

PR・アフィリエイト広告

MINISFORUM(ミニPC)

MINISFORUM(ミニPC)の公式広告バナーMINISFORUM公式ストアを見る ↗
PR本紙発行元 EmplifAI