Semantic Predictive Control for Explainable and Efficient Policy Learning

Pan, Xinlei; Chen, Xiangyu; Cai, Qi-Zhi; Canny, John; Yu, Fisher

doi:10.1109/icra.2019.8794437

Cited by 15 publications

(16 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The idea is to train a network to predict high level information such as semantic segmentation maps, distance to center of the lane, traffic light state etc... This prediction can then be used in several ways, either by a classic controller as in Sauer et al [31], either as auxiliary loss helping to find better features to the main imitative task loss as in Mehta et al [23] or also in a model-based RL approach as in the really recent work of Pan et al [26] while also providing some interpretable feedback on how the decision was taken.…”

Section: Introductionmentioning

confidence: 99%

End-to-End Model-Free Reinforcement Learning for Urban Driving Using Implicit Affordances

Toromanoff

Wirbel

Moutarde

2020

2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

140

132

View full text Add to dashboard Cite

Reinforcement Learning (RL) aims at learning an optimal behavior policy from its own experiments and not rulebased control methods. However, there is no RL algorithm yet capable of handling a task as difficult as urban driving. We present a novel technique, coined implicit affordances, to effectively leverage RL for urban driving thus including lane keeping, pedestrians and vehicles avoidance, and traffic light detection. To our knowledge we are the first to present a successful RL agent handling such a complex task especially regarding the traffic light detection. Furthermore, we have demonstrated the effectiveness of our method by winning the Camera Only track of the CARLA challenge.

show abstract

Section: Introductionmentioning

confidence: 99%

End-to-End Model-Free Reinforcement Learning for Urban Driving Using Implicit Affordances

Toromanoff

Wirbel

Moutarde

2020

2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

140

132

View full text Add to dashboard Cite

show abstract

“…Looking at autonomous vehicles as an example, Pan et al ( 2019 ), contributed Semantic Predictive Control (SPC) which learns to “predict the visual semantics of future states and possible events based upon visual inputs and an inferred sequence of future actions” (p. 3203). Visual semantics in this case refers to object detection, and the authors suggested that these predicted semantics can provide a visual explanation of the RL process.…”

Section: Discussionmentioning

confidence: 99%

“…Pan et al ( 2019 ) as previously described provided visual explanations in the form of object detection.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Explainable AI and Reinforcement Learning—A Systematic Review of Current Approaches and Trends

Wells

Bednarz

2021

Front. Artif. Intell.

103

View full text Add to dashboard Cite

Research into Explainable Artificial Intelligence (XAI) has been increasing in recent years as a response to the need for increased transparency and trust in AI. This is particularly important as AI is used in sensitive domains with societal, ethical, and safety implications. Work in XAI has primarily focused on Machine Learning (ML) for classification, decision, or action, with detailed systematic reviews already undertaken. This review looks to explore current approaches and limitations for XAI in the area of Reinforcement Learning (RL). From 520 search results, 25 studies (including 5 snowball sampled) are reviewed, highlighting visualization, query-based explanations, policy summarization, human-in-the-loop collaboration, and verification as trends in this area. Limitations in the studies are presented, particularly a lack of user studies, and the prevalence of toy-examples and difficulties providing understandable explanations. Areas for future study are identified, including immersive visualization, and symbolic representation.

show abstract

“…Over the past years, machine vision and decision-making systems have approached or surpassed human-level performance in vision and robotics applications due to the emergence of deep learning methods [16]. However, learning drive autonomously remains one of the most desirable but notoriously challenging problems, with high requirements on reliability, explainability, and data efficiency [26], [23], [11].…”

Section: Introductionmentioning

confidence: 99%

Instance-Aware Predictive Navigation in Multi-Agent Environments

Cao¹,

Wang²,

Darrell³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

In this work, we aim to achieve efficient endto-end learning of driving policies in dynamic multi-agent environments. Predicting and anticipating future events at the object level are critical for making informed driving decisions. We propose an Instance-Aware Predictive Control (IPC) approach, which forecasts interactions between agents as well as future scene structures. We adopt a novel multi-instance event prediction module to estimate the possible interaction among agents in the ego-centric view, conditioned on the selected action sequence of the ego-vehicle. To decide the action at each step, we seek the action sequence that can lead to safe future states based on the prediction module outputs by repeatedly sampling likely action sequences. We design a sequential action sampling strategy to better leverage predicted states on both scene-level and instance-level. Our method establishes a new state of the art in the challenging CARLA multi-agent driving simulation environments without expert demonstration, giving better explainability and sample efficiency.

show abstract

Semantic Predictive Control for Explainable and Efficient Policy Learning

Cited by 15 publications

References 19 publications

End-to-End Model-Free Reinforcement Learning for Urban Driving Using Implicit Affordances

End-to-End Model-Free Reinforcement Learning for Urban Driving Using Implicit Affordances

Explainable AI and Reinforcement Learning—A Systematic Review of Current Approaches and Trends

Instance-Aware Predictive Navigation in Multi-Agent Environments

Contact Info

Product

Resources

About