Compression has emerged as one of the essential deep learning research topics, especially for the edge devices that have limited computation power and storage capacity. Among the main compression techniques, low-rank compression via matrix factorization has been known to have two problems. First, an extensive tuning is required. Second, the resulting compression performance is typically not impressive. In this work, we propose a low-rank compression method that utilizes a modified beam-search for an automatic rank selection and a modified stable rank for a compression-friendly training. The resulting BSR (Beam-search and Stable Rank) algorithm requires only a single hyperparameter to be tuned for the desired compression ratio. The performance of BSR in terms of accuracy and compression ratio trade-off curve turns out to be superior to the previously known low-rank compression methods. Furthermore, BSR can perform on par with or better than the state-of-the-art structured pruning methods. As with pruning, BSR can be easily combined with quantization for an additional compression.

Hi everyone, this will be the second instalment in my tutorial series for building a chess engine. This lesson will focus on building an AI agent that we can play. This lesson is going to be more technical than part 1, so please bear with me. I try to supply both equations and diagrams to help make things a little easier. Now that we have finished building our chess game, we can begin designing an AI that plays it.

In this research note we show that a simple justification system can be used to teach humans non-trivial strategies of the Olympic sport of curling. This is achieved by justifying the decisions of Kernel Regression UCT (KR-UCT), a tree search algorithm that derives curling strategies by playing the game with itself. Given an action returned by KR-UCT and the expected outcome of that action, we use a decision tree to produce a counterfactual justification of KR-UCT's decision. The system samples other possible outcomes and selects for presentation the outcomes that are most similar to the expected outcome in terms of visual features and most different in terms of expected end-game value. A user study with 122 people shows that the participants who had access to the justifications produced by our system achieved much higher scores in a curling test than those who only observed the decision made by KR-UCT and those with access to the justifications of a baseline system. This is, to the best of our knowledge, the first work showing that a justification system is able to teach humans non-trivial strategies learned by an algorithm operating in self play.

The quality of mesh generation has long been considered a vital aspect in providing engineers with reliable simulation results throughout the history of the Finite Element Method (FEM). The element extraction method, which is currently the most robust method, is used in business software. However, in order to speed up extraction, the approach is done by finding the next element that optimizes a target function, which can result in local mesh of bad quality after many time steps. We provide TreeMesh, a method that uses this method in conjunction with reinforcement learning (also possible with supervised learning) and a novel Monte-Carlo tree search (MCTS) (Coulom(2006), Kocsis and Szepesv\'ari(2006), Browne et~al.(2012)). The algorithm is based on a previously proposed approach (Pan et~al.(2021)). After making many improvements on DRL (algorithm, state-action-reward setting) and adding a MCTS, it outperforms the former work on the same boundary. Furthermore, using tree search, our program reveals much preponderance on seed-density-changing boundaries, which is common on thin-film materials.

Artificial Intelligence, when amalgamated with games makes the ideal structure for research and advancing the field. Multi-agent games have multiple controls for each agent which generates huge amounts of data while increasing search complexity. Thus, we need advanced search methods to find a solution and create an artificially intelligent agent. In this paper, we propose our novel Evolutionary Monte Carlo Tree Search (FEMCTS) agent which borrows ideas from Evolutionary Algorthims (EA) and Monte Carlo Tree Search (MCTS) to play the game of Pommerman. It outperforms Rolling Horizon Evolutionary Algorithm (RHEA) significantly in high observability settings and performs almost as well as MCTS for most game seeds, outperforming it in some cases.

In this work, we introduce a novel workflow that couples robotics to machine-learning for efficient optimization of a non-aqueous battery electrolyte. A custom-built automated experiment named "Clio" is coupled to Dragonfly - a Bayesian optimization-based experiment planner. Clio autonomously optimizes electrolyte conductivity over a single-salt, ternary solvent design space. Using this workflow, we identify 6 fast-charging electrolytes in 2 work-days and 42 experiments (compared with 60 days using exhaustive search of the 1000 possible candidates, or 6 days assuming only 10% of candidates are evaluated). Our method finds the highest reported conductivity electrolyte in a design space heavily explored by previous literature, converging on a high-conductivity mixture that demonstrates subtle electrolyte chemical physics.

Adversarial attacks based on randomized search schemes have obtained state-of-the-art results in black-box robustness evaluation recently. However, as we demonstrate in this work, their efficiency in different query budget regimes depends on manual design and heuristic tuning of the underlying proposal distributions. We study how this issue can be addressed by adapting the proposal distribution online based on the information obtained during the attack. We consider Square Attack, which is a state-of-the-art score-based black-box attack, and demonstrate how its performance can be improved by a learned controller that adjusts the parameters of the proposal distribution online during the attack. We train the controller using gradient-based end-to-end training on a CIFAR10 model with white box access. We demonstrate that plugging the learned controller into the attack consistently improves its black-box robustness estimate in different query regimes by up to 20% for a wide range of different models with black-box access. We further show that the learned adaptation principle transfers well to the other data distributions such as CIFAR100 or ImageNet and to the targeted attack setting.

Active inference is a state-of-the-art framework for modelling the brain that explains a wide range of mechanisms such as habit formation, dopaminergic discharge and curiosity. However, recent implementations suffer from an exponential (space and time) complexity class when computing the prior over all the possible policies up to the time horizon. Fountas et al. (2020) used Monte Carlo tree search to address this problem, leading to very good results in two different tasks. Additionally, Champion et al. (2021a) proposed a tree search approach based on structure learning. This was enabled by the development of a variational message passing approach to active inference (Champion et al., 2021b), which enables compositional construction of Bayesian networks for active inference. However, this message passing tree search approach, which we call branching-time active inference (BTAI), has never been tested empirically. In this paper, we present an experimental study of the approach (Champion et al., 2021a) in the context of a maze solving agent. In this context, we show that both improved prior preferences and deeper search help mitigate the vulnerability to local minima. Then, we compare BTAI to standard active inference (AI) on a graph navigation task. We show that for small graphs, both BTAI and AI successfully solve the task. For larger graphs, AI exhibits an exponential (space) complexity class, making the approach intractable. However, BTAI explores the space of policies more efficiently, successfully scaling to larger graphs.

Over the last 10 to 15 years, active inference has helped to explain various brain mechanisms from habit formation to dopaminergic discharge and even modelling curiosity. However, the current implementations suffer from an exponential (space and time) complexity class when computing the prior over all the possible policies up to the time-horizon. Fountas et al (2020) used Monte Carlo tree search to address this problem, leading to impressive results in two different tasks. In this paper, we present an alternative framework that aims to unify tree search and active inference by casting planning as a structure learning problem. Two tree search algorithms are then presented. The first propagates the expected free energy forward in time (i.e., towards the leaves), while the second propagates it backward (i.e., towards the root). Then, we demonstrate that forward and backward propagations are related to active inference and sophisticated inference, respectively, thereby clarifying the differences between those two planning strategies.

Player 1 first starts with a word in mind that is at least 4 letters long, e.g. He plays the first letter of the word, i.e. 'm'.