In recent years, a wide range of investment models have been created using artificial intelligence. Automatic trading by artificial intelligence can expand the range of trading methods, such as by conferring the ability to operate 24 hours a day and the ability to trade with high frequency. Automatic trading can also be expected to trade with more information than is available to humans if it can sufficiently consider past data. In this paper, we propose an investment agent based on a deep reinforcement learning model, which is an artificial intelligence model. The model considers the transaction costs involved in actual trading and creates a framework for trading over a long period of time so that it can make a large profit on a single trade. In doing so, it can maximize the profit while keeping transaction costs low. In addition, in consideration of actual operations, we use online learning so that the system can continue to learn by constantly updating the latest online data instead of learning with static data. This makes it possible to trade in non-stationary financial markets by always incorporating current market trend information.

The shortcomings of maximum likelihood estimation in the context of model-based reinforcement learning have been highlighted by an increasing number of papers. When the model class is misspecified or has a limited representational capacity, model parameters with high likelihood might not necessarily result in high performance of the agent on a downstream control task. To alleviate this problem, we propose an end-to-end approach for model learning which directly optimizes the expected returns using implicit differentiation. We treat a value function that satisfies the Bellman optimality operator induced by the model as an implicit function of model parameters and show how to differentiate the function. We provide theoretical and empirical evidence highlighting the benefits of our approach in the model misspecification regime compared to likelihood-based methods.

Continual Learning (CL) considers the problem of training an agent sequentially on a set of tasks while seeking to retain performance on all previous tasks. A key challenge in CL is catastrophic forgetting, which arises when performance on a previously mastered task is reduced when learning a new task. While a variety of methods exist to combat forgetting, in some cases tasks are fundamentally incompatible with each other and thus cannot be learnt by a single policy. This can occur, in reinforcement learning (RL) when an agent may be rewarded for achieving different goals from the same observation. In this paper we formalize this ``interference'' as distinct from the problem of forgetting. We show that existing CL methods based on single neural network predictors with shared replay buffers fail in the presence of interference. Instead, we propose a simple method, OWL, to address this challenge. OWL learns a factorized policy, using shared feature extraction layers, but separate heads, each specializing on a new task. The separate heads in OWL are used to prevent interference. At test time, we formulate policy selection as a multi-armed bandit problem, and show it is possible to select the best policy for an unknown task using feedback from the environment. The use of bandit algorithms allows the OWL agent to constructively re-use different continually learnt policies at different times during an episode. We show in multiple RL environments that existing replay based CL methods fail, while OWL is able to achieve close to optimal performance when training sequentially.

We present an end-to-end framework for the Assignment Problem with multiple tasks mapped to a group of workers, using reinforcement learning while preserving many constraints. Tasks and workers have time constraints and there is a cost associated with assigning a worker to a task. Each worker can perform multiple tasks until it exhausts its allowed time units (capacity). We train a reinforcement learning agent to find near optimal solutions to the problem by minimizing total cost associated with the assignments while maintaining hard constraints. We use proximal policy optimization to optimize model parameters. The model generates a sequence of actions in real-time which correspond to task assignment to workers, without having to retrain for changes in the dynamic state of the environment. In our problem setting reward is computed as negative of the assignment cost. We also demonstrate our results on bin packing and capacitated vehicle routing problem, using the same framework. Our results outperform Google OR-Tools using MIP and CP-SAT solvers with large problem instances, in terms of solution quality and computation time.

We present a general optimization framework for emergent belief-state representation without any supervision. We employed the common configuration of multiagent reinforcement learning and communication to improve exploration coverage over an environment by leveraging the knowledge of each agent. In this paper, we obtained that recurrent neural nets (RNNs) with shared weights are highly biased in partially observable environments because of their noncooperativity. To address this, we designated an unbiased version of self-play via mechanism design, also known as reverse game theory, to clarify unbiased knowledge at the Bayesian Nash equilibrium. The key idea is to add imaginary rewards using the peer prediction mechanism (Miller et al., 2005), i.e., a mechanism for mutually criticizing information in a decentralized environment. Numerical analyses, including StarCraft exploration tasks with up to 20 agents and off-the-shelf RNNs, demonstrate the state-of-the-art performance.

Imitation learning and instruction-following are two common approaches to communicate a user's intent to a learning agent. However, as the complexity of tasks grows, it could be beneficial to use both demonstrations and language to communicate with an agent. In this work, we propose a novel setting where an agent is given both a demonstration and a description, and must combine information from both the modalities. Specifically, given a demonstration for a task (the source task), and a natural language description of the differences between the demonstrated task and a related but different task (the target task), our goal is to train an agent to complete the target task in a zero-shot setting, that is, without any demonstrations for the target task. To this end, we introduce Language-Aided Reward and Value Adaptation (LARVA) which, given a source demonstration and a linguistic description of how the target task differs, learns to output a reward / value function that accurately describes the target task. Our experiments show that on a diverse set of adaptations, our approach is able to complete more than 95% of target tasks when using template-based descriptions, and more than 70% when using free-form natural language.

In reinforcement learning, experience replay stores past samples for further reuse. Prioritized sampling is a promising technique to better utilize these samples. Previous criteria of prioritization include TD error, recentness and corrective feedback, which are mostly heuristically designed. In this work, we start from the regret minimization objective, and obtain an optimal prioritization strategy for Bellman update that can directly maximize the return of the policy. The theory suggests that data with higher hindsight TD error, better on-policiness and more accurate Q value should be assigned with higher weights during sampling. Thus most previous criteria only consider this strategy partially. We not only provide theoretical justifications for previous criteria, but also propose two new methods to compute the prioritization weight, namely ReMERN and ReMERT. ReMERN learns an error network, while ReMERT exploits the temporal ordering of states. Both methods outperform previous prioritized sampling algorithms in challenging RL benchmarks, including MuJoCo, Atari and Meta-World.

We provide a framework for accelerating reinforcement learning (RL) algorithms by heuristics constructed from domain knowledge or offline data. Tabula rasa RL algorithms require environment interactions or computation that scales with the horizon of the sequential decision-making task. Using our framework, we show how heuristic-guided RL induces a much shorter-horizon subproblem that provably solves the original task. Our framework can be viewed as a horizon-based regularization for controlling bias and variance in RL under a finite interaction budget. On the theoretical side, we characterize properties of a good heuristic and its impact on RL acceleration. In particular, we introduce the novel concept of an "improvable heuristic" -- a heuristic that allows an RL agent to extrapolate beyond its prior knowledge. On the empirical side, we instantiate our framework to accelerate several state-of-the-art algorithms in simulated robotic control tasks and procedurally generated games. Our framework complements the rich literature on warm-starting RL with expert demonstrations or exploratory datasets, and introduces a principled method for injecting prior knowledge into RL.

When solving two-player zero-sum games, multi-agent reinforcement learning (MARL) algorithms often create populations of agents where, at each iteration, a new agent is discovered as the best response to a mixture over the opponent population. Within such a process, the update rules of "who to compete with" (i.e., the opponent mixture) and "how to beat them" (i.e., finding best responses) are underpinned by manually developed game theoretical principles such as fictitious play and Double Oracle. In this paper we introduce a framework, LMAC, based on meta-gradient descent that automates the discovery of the update rule without explicit human design. Specifically, we parameterise the opponent selection module by neural networks and the best-response module by optimisation subroutines, and update their parameters solely via interaction with the game engine, where both players aim to minimise their exploitability. Surprisingly, even without human design, the discovered MARL algorithms achieve competitive or even better performance with the state-of-the-art population-based game solvers (e.g., PSRO) on Games of Skill, differentiable Lotto, non-transitive Mixture Games, Iterated Matching Pennies, and Kuhn Poker. Additionally, we show that LMAC is able to generalise from small games to large games, for example training on Kuhn Poker and outperforming PSRO on Leduc Poker. Our work inspires a promising future direction to discover general MARL algorithms solely from data.

Recent work in AI safety has highlighted that in sequential decision making, objectives are often underspecified or incomplete. This gives discretion to the acting agent to realize the stated objective in ways that may result in undesirable outcomes. We contend that to learn to act safely, a reinforcement learning (RL) agent should include contemplation of the impact of its actions on the wellbeing and agency of others in the environment, including other acting agents and reactive processes. We endow RL agents with the ability to contemplate such impact by augmenting their reward based on expectation of future return by others in the environment, providing different criteria for characterizing impact. We further endow these agents with the ability to differentially factor this impact into their decision making, manifesting behavior that ranges from self-centred to self-less, as demonstrated by experiments in gridworld environments.