We study the problem of learning an approximate equilibrium in the offline multi-agent reinforcement learning (MARL) setting. We introduce a structural assumption---the interaction rank---and establish that functions with low interaction rank are significantly more robust to distribution shift compared to general ones. Leveraging this observation, we demonstrate that utilizing function classes with low interaction rank, when combined with regularization and no-regret learning, admits decentralized, computationally and statistically efficient learning in cooperative and competitive offline MARL. Our theoretical results are complemented by experiments that showcase the potential of critic architectures with low interaction rank in offline MARL, contrasting with commonly used single-agent value decomposition architectures.

To date, the multi-objective optimization literature has mainly focused on conflicting objectives, studying the Pareto front, or requiring users to balance tradeoffs. Yet, in machine learning practice, there are many scenarios where such conflict does not take place. Recent findings from multi-task learning, reinforcement learning, and LLMs training show that diverse related tasks can enhance performance across objectives simultaneously. Despite this evidence, such phenomenon has not been examined from an optimization perspective. This leads to a lack of generic gradient-based methods that can scale to scenarios with a large number of related objectives. To address this gap, we introduce the Aligned Multi-Objective Optimization framework, propose new algorithms for this setting, and provide theoretical guarantees of its superior performance compared to naive approaches.

Online optimization of resource management for large-scale data centers and infrastructures to meet dynamic capacity reservation demands and various practical constraints (e.g., feasibility and robustness) is a very challenging problem. Mixed Integer Programming (MIP) approaches suffer from recognized limitations in such a dynamic environment, while learning-based approaches may face with prohibitively large state/action spaces. To this end, this paper presents a novel two-tiered online optimization to enable a learning-based Resource Allowance System (RAS). To solve optimal server-to-reservation assignment in RAS in an online fashion, the proposed solution leverages a reinforcement learning (RL) agent to make high-level decisions, e.g., how much resource to select from the Main Switch Boards (MSBs), and then a low-level Mixed Integer Linear Programming (MILP) solver to generate the local server-to-reservation mapping, conditioned on the RL decisions. We take into account fault tolerance, server movement minimization, and network affinity requirements and apply the proposed solution to large-scale RAS problems. To provide interpretability, we further train a decision tree model to explain the learned policies and to prune unreasonable corner cases at the low-level MILP solver, resulting in further performance improvement. Extensive evaluations show that our two-tiered solution outperforms baselines such as pure MIP solver by over 15% while delivering$100\times $speedup in computation.

Model-based reinforcement learning (RL) has shown great promise due to its sample efficiency, but still struggles with long-horizon sparse-reward tasks, especially in offline settings where the agent learns from a fixed dataset. We hypothesize that model-based RL agents struggle in these environments due to a lack of long-term planning capabilities, and that planning in a temporally abstract model of the environment can alleviate this issue. In this paper, we make two key contributions: 1) we introduce an offline model-based RL algorithm, IQL-TD-MPC, that extends the state- of-the-art Temporal Difference Learning for Model Predictive Control (TD-MPC) with Implicit Q-Learning (IQL); and 2) we propose to use IQL-TD-MPC as a Manager in a hierarchical setting with any off-the-shelf offline RL algorithm as a Worker. More specifically, we pre-train a temporally abstract IQL-TD-MPC Manager to predict "intent embeddings", which roughly correspond to subgoals, via planning. We show that augmenting state representations with intent embeddings generated by an IQL-TD-MPC manager significantly improves off-the-shelf offline RL agents' performance on some of the most challenging D4RL benchmark tasks. For instance, the offline RL algorithms AWAC, TD3-BC, DT, and CQL all get zero or near-zero normalized evaluation scores on the medium and large antmaze tasks, while our modification gives an average score over 40.

The online advertising market, with its thousands of auctions run per second, presents a daunting challenge for advertisers who wish to optimize their spend under a budget constraint. Thus, advertising platforms typically provide automated agents to their customers, which act on their behalf to bid for impression opportunities in real time at scale. Because these proxy agents are owned by the platform but use advertiser funds to operate, there is a strong practical need to balance reliability and explainability of the agent with optimizing power. We propose a generalizable approach to optimizing bidding policies in production environments by learning from real data using offline reinforcement learning. This approach can be used to optimize any differentiable base policy (practically, a heuristic policy based on principles which the advertiser can easily understand), and only requires data generated by the base policy itself. We use a hybrid agent architecture that combines arbitrary base policies with deep neural networks, where only the optimized base policy parameters are eventually deployed, and the neural network part is discarded after training. We demonstrate that such an architecture achieves statistically significant performance gains in both simulated and at-scale production bidding environments compared to the default production bidding policy. Our approach does not incur additional infrastructure, safety, or explainability costs, as it directly optimizes the parameters of existing production routines without necessarily replacing them with black box-style models like neural networks.

Neural Collapse (NC) was a recently discovered phenomenon that the output features and the classifier weights of the neural network converge to optimal geometric structures at the Terminal Phase of Training (TPT) under various losses. However, the relationship between these optimal structures at TPT and the classification performance remains elusive, especially in imbalanced learning. Even though it is noticed that fixing the classifier to an optimal structure can mitigate the minority collapse problem, the performance is still not comparable to the classical imbalanced learning methods with a learnable classifier. In this work, we find that the optimal structure can be designed to represent a better classification rule, and thus achieve better performance. In particular, we justify that, to achieve better classification, the features from the minor classes should align with more directions. This justification then yields a decision rule called the Generalized Classification Rule (GCR) and we also term these directions as the centers of the classes. Then we study the NC under an MSE-type loss via the Unconstrained Features Model (UFM) framework where (1) the features from a class tend to collapse to the mean of the corresponding centers of that class (named Neural Collapse to Multiple Centers (NCMC)) at the global optimum, and (2) the original classifier approximates a surrogate to GCR when NCMC occurs. Based on the analysis, we develop a strategy for determining the number of centers and propose a Cosine Loss function for the fixed classifier that induces NCMC. Our experiments have shown that the Cosine Loss can induce NCMC and has performance on long-tail classification comparable to the classical imbalanced learning methods.

Reinforcement learning (RL) is a versatile framework for optimizing long-term goals. Although many real-world problems can be formalized with RL, learning and deploying a performant RL policy requires a system designed to address several important challenges, including the exploration-exploitation dilemma, partial observability, dynamic action spaces, and safety concerns. While the importance of these challenges has been well recognized, existing open-source RL libraries do not explicitly address them. This paper introduces Pearl, a Production-Ready RL software package designed to embrace these challenges in a modular way. In addition to presenting benchmarking results, we also highlight examples of Pearl's ongoing industry adoption to demonstrate its advantages for production use cases. Pearl is open sourced on GitHub at github.com/facebookresearch/pearl and its official website is pearlagent.github.io.

Abstract Many Internet applications adopt real-time bidding mechanisms to ensure different services (types of content) are shown to the users through fair competitions. The service offering the highest bid price gets the content slot to present a list of items in its candidate pool. Through user interactions with the recommended items, the service obtains the desired engagement activities. We propose a contextual-bandit framework to jointly optimize the price to bid for the slot and the order to rank its candidates for a given service in this type of recommendation systems. Our method can take as input any feature that describes the user and the candidates, including the outputs of other machine learning models. We train reinforcement learning policies using deep neural networks, and compute top-K Gaussian propensity scores to exclude the variance in the gradients caused by randomness unrelated to the reward. This setup further facilitates us to automatically find accurate reward functions that trade off between budget spending and user engagements. In online A/B experiments on two major services of Facebook Home Feed, Groups You Should Join and Friend Requests, our method statistically significantly boosted the number of groups joined by 14.7%, the number of friend requests accepted by 7.0%, and the number of daily active Facebook users by about 1 million, against strong hand-tuned baselines that have been iterated in production over years.

High-quality recommender systems ought to deliver both innovative and relevant content through effective and exploratory interactions with users. Yet, supervised learning-based neural networks, which form the backbone of many existing recommender systems, only leverage recognized user interests, falling short when it comes to efficiently uncovering unknown user preferences. While there has been some progress with neural contextual bandit algorithms towards enabling online exploration through neural networks, their onerous computational demands hinder widespread adoption in real-world recommender systems. In this work, we propose a scalable sample-efficient neural contextual bandit algorithm for recommender systems. To do this, we design an epistemic neural network architecture, Epistemic Neural Recommendation (ENR), that enables Thompson sampling at a large scale. In two distinct large-scale experiments with real-world tasks, ENR significantly boosts click-through rates and user ratings by at least 9% and 6% respectively compared to state-of-the-art neural contextual bandit algorithms. Furthermore, it achieves equivalent performance with at least 29% fewer user interactions compared to the best-performing baseline algorithm. Remarkably, while accomplishing these improvements, ENR demands orders of magnitude fewer computational resources than neural contextual bandit baseline algorithms.

Auction-based recommender systems are prevalent in online advertising platforms, but they are typically optimized to allocate recommendation slots based on immediate expected return metrics, neglecting the downstream effects of recommendations on user behavior. In this study, we employ reinforcement learning to optimize for long-term return metrics in an auction-based recommender system. Utilizing temporal difference learning, a fundamental reinforcement learning algorithm, we implement a one-step policy improvement approach that biases the system towards recommendations with higher long-term user engagement metrics. This optimizes value over long horizons while maintaining compatibility with the auction framework. Our approach is grounded in dynamic programming ideas which show that our method provably improves upon the existing auction-based base policy. Through an online A/B test conducted on an auction-based recommender system which handles billions of impressions and users daily, we empirically establish that our proposed method outperforms the current production system in terms of long-term user engagement metrics.

Modern recommendation systems ought to benefit by probing for and learning from delayed feedback. Research has tended to focus on learning from a user’s response to a single recommendation. Such work, which leverages methods of supervised and bandit learning, forgoes learning from the user’s subsequent behavior. Where past work has aimed to learn from subsequent behavior, there has been a lack of effective methods for probing to elicit informative delayed feedback. Effective exploration through probing for delayed feedback becomes particularly challenging when rewards are sparse. To address this, we develop deep exploration methods for recommendation systems. In particular, we formulate recommendation as a sequential decision problem and demonstrate benefits of deep exploration over single-step exploration. Our experiments are carried out with high-fidelity industrial-grade simulators and establish large improvements over existing algorithms.

Multi-agent safe systems have become an increasingly important area of study as we can now easily have multiple AI-powered systems operating together. In such settings, we need to ensure the safety of not only each individual agent, but also the overall system. In this paper, we introduce a novel multi-agent safe learning algorithm that enables decentralized safe navigation when there are multiple different agents in the environment. This algorithm makes mild assumptions about other agents and is trained in a decentralized fashion, i.e. with very little prior knowledge about other agents' policies. Experiments show our algorithm performs well with the robots running other algorithms when optimizing various objectives.