On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,878 papers · 148 categories

Trend · papers per month

count % of corpus

3 results for “AI2-THOR”

Vision-based Navigation Using Deep Reinforcement Learningcs.RO

Deep RL improves robot navigation in images.

problem Applying deep RL to visual navigation in realistic environments.

method Extended A2C algorithm with auxiliary tasks for segmentation, depth prediction, and target prediction.

result Method outperforms state-of-the-art visual navigation methods.

Deep Reinforcement Learning Robot Navigation Vision-based Image Recognition AI2-THOR

2019-08-08abs ↗pdf ↗

Dynamic Value Estimation for Single-Task Multi-Scene Reinforcement Learningcs.LG

Improves reinforcement learning agent's scene-specific value function.

problem High variance in samples for policy gradient computations in multi-scene environments.

method Proposes dynamic value estimation (DVE) for multiple MDPs, clustering value functions across scenes.

result Lower sample variance and more accurate scene-specific value function estimates.

reinforcement learning multi-scene value estimation policy gradient Markov Decision Process

2020-05-25abs ↗pdf ↗

Improving Target-driven Visual Navigation with Attention on 3D Spatial Relationshipscs.CV

Improves AI agents' 3D navigation by learning from failures and 3D spatial relationships.

problem Challenges in data efficiency, obstacle avoidance, and generalization in 3D visual navigation.

method Incorporates attention on 3D spatial relationships and a target skill extension module into DRL framework.

result Significantly improves navigation performance and generalization across targets and scenes.

3D navigation attention target skill extension graph convolutional network reinforcement learning

2020-04-29abs ↗pdf ↗