Ranking is a fundamental and widely studied problem in scenarios such as search, advertising, and recommendation. However, joint optimization for multi-scenario ranking, which aims to improve the overall performance of several ranking strategies in different scenarios, is rather untouched. Separately optimizing each individual strategy has two limitations. The first one is lack of collaboration between scenarios meaning that each strategy maximizes its own objective but ignores the goals of other strategies, leading to a sub-optimal overall performance. The second limitation is the inability of modeling the correlation between scenarios meaning that independent optimization in one scenario only uses its own user data but ignores the context in other scenarios. In this paper, we formulate multi-scenario ranking as a fully cooperative, partially observable, multi-agent sequential decision problem. We propose a novel model named Multi-Agent Recurrent Deterministic Policy Gradient (MA-RDPG) which has a communication component for passing messages, several private actors (agents) for making actions for ranking, and a centralized critic for evaluating the overall performance of the co-working actors. Each scenario is treated as an agent (actor). Agents collaborate with each other by sharing a global action-value function (the critic) and passing messages that encodes historical information across scenarios. The model is evaluated with online settings on a large E-commerce platform. Results show that the proposed model exhibits significant improvements against baselines in terms of the overall performance.

In Multi-Agent Reinforcement Learning, independent cooperative learners must overcome a number of pathologies in order to learn optimal joint policies. These pathologies include action-shadowing, stochasticity, the moving target and alter-exploration problems (Matignon, Laurent, and Le Fort-Piat 2012; Wei and Luke 2016). Numerous methods have been proposed to address these pathologies, but evaluations are predominately conducted in repeated strategic-form games and stochastic games consisting of only a small number of state transitions. This raises the question of the scalability of the methods to complex, temporally extended, partially observable domains with stochastic transitions and rewards. In this paper we study such complex settings, which require reasoning over long time horizons and confront agents with the curse of dimensionality. To deal with the dimensionality, we adopt a Multi-Agent Deep Reinforcement Learning (MA-DRL) approach. We find that when the agents have to make critical decisions in seclusion, existing methods succumb to a combination of relative overgeneralisation (a type of action shadowing), the alter-exploration problem, and the stochasticity. To address these pathologies we introduce expanding negative update intervals that enable independent learners to establish the near-optimal average utility values for higher-level strategies while largely discarding transitions from episodes that result in mis-coordination. We evaluate Negative Update Intervals Double-DQN (NUI-DDQN) within a temporally extended Climb Game, a normal form game which has frequently been used to study relative over-generalisation and other pathologies. We show that NUI-DDQN can converge towards optimal joint-policies in deterministic and stochastic reward settings, overcoming relative-overgeneralisation and the alter-exploration problem while mitigating the moving target problem.

We address item relocation problems in graphs in this paper. We assume items placed in vertices of an undirected graph with at most one item per vertex. Items can be moved across edges while various constraints depending on the type of relocation problem must be satisfied. We introduce a general problem formulation that encompasses known types of item relocation problems such as multi-agent path finding (MAPF) and token swapping (TSWAP). In this formulation we express two new types of relocation problems derived from token swapping that we call token rotation (TROT) and token permutation (TPERM). Our solving approach for item relocation combines satisfiability modulo theory (SMT) with conflict-based search (CBS). We interpret CBS in the SMT framework where we start with the basic model and refine the model with a collision resolution constraint whenever a collision between items occurs in the current solution. The key difference between the standard CBS and our SMT-based modification of CBS (SMT-CBS) is that the standard CBS branches the search to resolve the collision while in SMT-CBS we iteratively add a single disjunctive collision resolution constraint. Experimental evaluation on several benchmarks shows that the SMT-CBS algorithm significantly outperforms the standard CBS. We also compared SMT-CBS with a modification of the SAT-based MDD-SAT solver that uses an eager modeling of item relocation in which all potential collisions are eliminated by constrains in advance. Experiments show that lazy approach in SMT-CBS produce fewer constraint than MDD-SAT and also achieves faster solving run-times.

Specialization and hierarchical organization are important features of efficient collaboration in economical, artificial, and biological systems. Here, we investigate the hypothesis that both features can be explained by the fact that each entity of such a system is limited in a certain way. We propose an information-theoretic approach based on a Free Energy principle, in order to computationally analyze systems of bounded rational agents that deal with such limitations optimally. We find that specialization allows to focus on fewer tasks, thus leading to a more efficient execution, but in turn requires coordination in hierarchical structures of specialized experts and coordinating units. Our results suggest that hierarchical architectures of specialized units at lower levels that are coordinated by units at higher levels are optimal, given that each unit's information-processing capability is limited and conforms to constraints on complexity costs.

Wagering mechanisms are one-shot betting mechanisms that elicit agents' predictions of an event. For deterministic wagering mechanisms, an existing impossibility result has shown incompatibility of some desirable theoretical properties. In particular, Pareto optimality (no profitable side bet before allocation) can not be achieved together with weak incentive compatibility, weak budget balance and individual rationality. In this paper, we expand the design space of wagering mechanisms to allow randomization and ask whether there are randomized wagering mechanisms that can achieve all previously considered desirable properties, including Pareto optimality. We answer this question positively with two classes of randomized wagering mechanisms: i) one simple randomized lottery-type implementation of existing deterministic wagering mechanisms, and ii) another family of simple and randomized wagering mechanisms which we call surrogate wagering mechanisms, which are robust to noisy ground truth. This family of mechanisms builds on the idea of learning with noisy labels (Natarajan et al. 2013) as well as a recent extension of this idea to the information elicitation without verification setting (Liu and Chen 2018). We show that a broad family of randomized wagering mechanisms satisfy all desirable theoretical properties.

Autonomous robotics and artificial intelligence techniques can be used to support human personnel in the event of critical incidents. These incidents can pose great danger to human life. Some examples of such assistance include: multi-robot surveying of the scene; collection of sensor data and scene imagery, real-time risk assessment and analysis; object identification and anomaly detection; and retrieval of relevant supporting documentation such as standard operating procedures (SOPs). These incidents, although often rare, can involve chemical, biological, radiological/nuclear or explosive (CBRNE) substances and can be of high consequence. Real-world training and deployment of these systems can be costly and sometimes not feasible. For this reason, we have developed a realistic 3D model of a CBRNE scenario to act as a testbed for an initial set of assisting AI tools that we have developed.

There are multiple notions of coalitional responsibility. The focus of this paper is on the blameworthiness defined through the principle of alternative possibilities: a coalition is blamable for a statement if the statement is true, but the coalition had a strategy to prevent it. The main technical result is a sound and complete bimodal logical system that describes properties of blameworthiness in one-shot games.

We propose CM3, a new deep reinforcement learning method for cooperative multi-agent problems where agents must coordinate for joint success in achieving different individual goals. We restructure multi-agent learning into a two-stage curriculum, consisting of a single-agent stage for learning to accomplish individual tasks, followed by a multi-agent stage for learning to cooperate in the presence of other agents. These two stages are bridged by modular augmentation of neural network policy and value functions. We further adapt the actor-critic framework to this curriculum by formulating local and global views of the policy gradient and learning via a double critic, consisting of a decentralized value function and a centralized action-value function. We evaluated CM3 on a new high-dimensional multi-agent environment with sparse rewards: negotiating lane changes among multiple autonomous vehicles in the Simulation of Urban Mobility (SUMO) traffic simulator. Detailed ablation experiments show the positive contribution of each component in CM3, and the overall synthesis converges significantly faster to higher performance policies than existing cooperative multi-agent methods.

We discuss the role of coordination as a direct learning objective in multi-agent reinforcement learning (MARL) domains. To this end, we present a novel means of quantifying coordination in multi-agent systems, and discuss the implications of using such a measure to optimize coordinated agent policies. This concept has important implications for adversary-aware RL, which we take to be a sub-domain of multi-agent learning.

Unity Technologies on Monday released version 0.5 of its ML-Agents toolkit to make its Unity 3D game development platform better suited for developing and training autonomous agent code via machine learning. Initially rolled out a year ago in beta, version 0.5 comes with a few improvements. There's a wrapper for Gym (a toolkit for developing and testing reinforcement learning algorithms), support for letting agents make multiple action selections at once and for preventing agents from taking certain actions, and a refurbished set of environments called Marathon Environments. In these virtual spaces, AI researchers can teach software agents to perform certain tasks by rewarding them for correct actions. This sort of reinforcement learning can be limited to digital environments like video games or mapped to software-driven machines in the real world. Through its latest code update, Unity is making the case for Unity 3D as a key tool for AI research, a goal that company code boffins describe in a preprint paper titled, "Unity: A General Platform for Intelligent Agents."