• Corpus ID: 246015566

Detecting danger in gridworlds using Gromov's Link Condition

  title={Detecting danger in gridworlds using Gromov's Link Condition},
  author={Thomas F Burns and Robert Tang},
Gridworlds have been long-utilised in AI research, particularly in reinforcement learning, as they provide simple yet scalable models for many real-world applications such as robot navigation, emergent behaviour, and operations research. We initiate a study of gridworlds using the mathematical framework of reconfigurable systems and state complexes due to Abrams, Ghrist & Peterson. State complexes represent all possible configurations of a system as a single geometric space, thus making them… 

Simplicial Hopfield networks

Surprisingly, even when connections are limited to a small random subset of equivalent size to an all-pairwise network, these networks still outperform their pairwise counterparts.

Acquisition of chess knowledge in AlphaZero

Remarkably, there are many strong correspondences between human concepts and AlphaZero’s representations that emerge during training, even though none of these concepts were initially present in the network.

A space of goals: the cognitive geometry of informationally bounded agents

This paper considers geometry in terms of travel between states within a world by incorporating information processing costs with the appropriate spatial distances, and approximates the usual geometric property that, travelling from a start to a goal along a geodesic, not only the goal, but all intermediate points are equally visited at optimal cost from the start.

On the geometry of generalization and memorization in deep neural networks

Analytical analysis shows that networks avoid memorization early in training because close to initialization, the gradient contribution from permuted examples are small, and this provides quantitative evidence for the structure of memorization across layers of a deep neural network, the drivers for such structure, and its connection to manifold geometric properties.

Flatland Competition 2020: MAPF and MARL for Efficient Train Coordination on a Grid World

The Flatland competition aimed at finding novel approaches to solve the vehicle re-scheduling problem (VRSP), and top submissions used graph representations of the environment to construct tree-based observations.

Quantitative Performance Assessment of CNN Units via Topological Entropy Calculation

This work proposes a novel method for quantitatively clarifying the status of single unit in CNN using algebraic topological tools and shows that feature entropy decreases as the layer goes deeper and shares almost simultaneous trend with loss during training.

Toroidal topology of population activity in grid cells

Simultaneous recordings from hundreds of grid cells in rats, combined with topological data analysis, show that network activity in grid cells resides on a toroidal manifold that is invariant across environments and brain states.

Learning Safe Multi-Agent Control with Decentralized Neural Barrier Certificates

This work proposes a novel joint-learning framework that can be implemented in a decentralized fashion, with generalization guarantees for certain function classes, and improves the scalability by incorporating neural network architectures that are invariant to the quantity and permutation of neighboring agents.

Deep reinforcement learning for drone navigation using sensor data

This paper describes a generic navigation algorithm that uses data from sensors on-board the drone to guide the drones to the site of the problem and uses the proximal policy optimisation deep reinforcement learning algorithm coupled with incremental curriculum learning and long short-term memory neural networks to implement it.

Distributed multi-robot collision avoidance via deep reinforcement learning for navigation in complex scenarios

A decentralized sensor-level collision-avoidance policy for multi-robot systems, which enables a robot to make effective progress in a crowd without getting stuck and has been successfully deployed on different types of physical robot platforms without tedious parameter tuning.