Brandon Fain

Zimeng Fan, Nianli Peng, Muhang Tian et al.

We study fair multi-objective reinforcement learning in which an agent must learn a policy that simultaneously achieves high reward on multiple dimensions of a vector-valued reward. Motivated by the fair resource allocation literature, we model this as an expected welfare maximization problem, for some nonlinear fair welfare function of the vector of long-term cumulative rewards. One canonical example of such a function is the Nash Social Welfare, or geometric mean, the log transform of which is also known as the Proportional Fairness objective. We show that even approximately optimal optimization of the expected Nash Social Welfare is computationally intractable even in the tabular case. Nevertheless, we provide a novel adaptation of Q-learning that combines nonlinear scalarized learning updates and non-stationary action selection to learn effective policies for optimizing nonlinear welfare functions. We show that our algorithm is provably convergent, and we demonstrate experimentally that our approach outperforms techniques based on linear scalarization, mixtures of optimal linear scalarizations, or stationary action selection for the Nash Social Welfare Objective.

Nianli Peng, Muhang Tian, Brandon Fain

We study multi-objective reinforcement learning with nonlinear preferences over trajectories. That is, we maximize the expected value of a nonlinear function over accumulated rewards (expected scalarized return or ESR) in a multi-objective Markov Decision Process (MOMDP). We derive an extended form of Bellman optimality for nonlinear optimization that explicitly considers time and current accumulated reward. Using this formulation, we describe an approximation algorithm for computing an approximately optimal non-stationary policy in pseudopolynomial time for smooth scalarization functions with a constant number of rewards. We prove the approximation analytically and demonstrate the algorithm experimentally, showing that there can be a substantial gap between the optimal policy computed by our algorithm and alternative baselines.

Chloe Qinyu Zhu, Rickard Stureborg, Brandon Fain

Language Representation Models (LRMs) trained with real-world data may capture and exacerbate undesired bias and cause unfair treatment of people in various demographic groups. Several techniques have been investigated for applying interventions to LRMs to remove bias in benchmark evaluations on, for example, word embeddings. However, the negative side effects of debiasing interventions are usually not revealed in the downstream tasks. We propose xGAP-DEBIAS, a set of evaluations on assessing the fairness of debiasing. In this work, We examine four debiasing techniques on a real-world text classification task and show that reducing biasing is at the cost of degrading performance for all demographic groups, including those the debiasing techniques aim to protect. We advocate that a debiasing technique should have good downstream performance with the constraint of ensuring no harm to the protected group.

Henry Bell, Lara Neubauer da Costa Schertel, Bochu Ding et al.

A crucial consideration when developing and deploying Large Language Models (LLMs) is the human values to which these models are aligned. In the constitutional framework of alignment models are aligned to a set of principles (the constitution) specified in natural language. However, it is unclear how to fairly determine this constitution with widespread stakeholder input. In this work we propose Grounded Constitutional AI (GCAI), a unified framework for generating constitutions of principles that are representative of both users' general expectations toward AI (general principles) and their interaction-time preferences (contextual principles). We extend the Inverse Constitutional AI (ICAI) approach to generate contextual principles from human preference annotation data by leveraging human-provided \textit{reasons} for their preferences. We supplement these contextual principles with general principles surfaced from user statements of \textit{values} regarding AI. We show that a constitution generated by GCAI is preferred by humans over one generated through ICAI both personally, and for widespread use in governing AI behavior. Additionally participants consider the GCAI constitution to be more morally grounded, coherent, and pluralistic.

Henry Bell, Caroline Zhang, Mohammed Mobasserul Haque et al.

The constitutional framework of alignment aims to align large language models (LLMs) with value-laden principles written in natural language (such as to avoid using biased language). Prior work has focused on parameter fine-tuning techniques, such as reinforcement learning from human feedback (RLHF), to instill these principles. However, these approaches are computationally demanding, require careful engineering and tuning, and often require difficult-to-obtain human annotation data. We propose \textsc{reflect}, an inference-time framework for constitutional alignment that does not require any training or data, providing a plug-and-play approach for aligning an instruction-tuned model to a set of principles. \textsc{reflect} operates entirely in-context, combining a (i) constitution-conditioned base response with post-generation (ii) self-evaluation, (iii)(a) self-critique, and (iii)(b) final revision. \textsc{reflect}'s technique of explicit in-context reasoning over principles during post-generation outperforms standard few-shot prompting and provides transparent reasoning traces. Our results demonstrate that \textsc{reflect} significantly improves LLM conformance to diverse and complex principles, including principles quite distinct from those emphasized in the model's original parameter fine-tuning, without sacrificing factual reasoning. \textsc{reflect} is particularly effective at reducing the rate of rare but significant violations of principles, thereby improving safety and robustness in the tail end of the distribution of generations. Finally, we show that \textsc{reflect} naturally generates useful training data for traditional parameter fine-tuning techniques, allowing for efficient scaling and the reduction of inference-time computational overhead in long-term deployment scenarios.

David Pujol, Yikai Wu, Brandon Fain et al.

Large organizations that collect data about populations (like the US Census Bureau) release summary statistics that are used by multiple stakeholders for resource allocation and policy making problems. These organizations are also legally required to protect the privacy of individuals from whom they collect data. Differential Privacy (DP) provides a solution to release useful summary data while preserving privacy. Most DP mechanisms are designed to answer a single set of queries. In reality, there are often multiple stakeholders that use a given data release and have overlapping but not-identical queries. This introduces a novel joint optimization problem in DP where the privacy budget must be shared among different analysts. We initiate study into the problem of DP query answering across multiple analysts. To capture the competing goals and priorities of multiple analysts, we formulate three desiderata that any mechanism should satisfy in this setting -- The Sharing Incentive, Non-Interference, and Adaptivity -- while still optimizing for overall error. We demonstrate how existing DP query answering mechanisms in the multi-analyst settings fail to satisfy at least one of the desiderata. We present novel DP algorithms that provably satisfy all our desiderata and empirically show that they incur low error on realistic tasks.

Xingyu Chen, Brandon Fain, Liang Lyu et al.

We extend the fair machine learning literature by considering the problem of proportional centroid clustering in a metric context. For clustering $n$ points with $k$ centers, we define fairness as proportionality to mean that any $n/k$ points are entitled to form their own cluster if there is another center that is closer in distance for all $n/k$ points. We seek clustering solutions to which there are no such justified complaints from any subsets of agents, without assuming any a priori notion of protected subsets. We present and analyze algorithms to efficiently compute, optimize, and audit proportional solutions. We conclude with an empirical examination of the tradeoff between proportional solutions and the $k$-means objective.

Brandon Fain, Ashish Goel, Kamesh Munagala et al.

We study social choice mechanisms in an implicit utilitarian framework with a metric constraint, where the goal is to minimize \textit{Distortion}, the worst case social cost of an ordinal mechanism relative to underlying cardinal utilities. We consider two additional desiderata: Constant sample complexity and Squared Distortion. Constant sample complexity means that the mechanism (potentially randomized) only uses a constant number of ordinal queries regardless of the number of voters and alternatives. Squared Distortion is a measure of variance of the Distortion of a randomized mechanism. Our primary contribution is the first social choice mechanism with constant sample complexity \textit{and} constant Squared Distortion (which also implies constant Distortion). We call the mechanism Random Referee, because it uses a random agent to compare two alternatives that are the favorites of two other random agents. We prove that the use of a comparison query is necessary: no mechanism that only elicits the top-k preferred alternatives of voters (for constant k) can have Squared Distortion that is sublinear in the number of alternatives. We also prove that unlike any top-k only mechanism, the Distortion of Random Referee meaningfully improves on benign metric spaces, using the Euclidean plane as a canonical example. Finally, among top-1 only mechanisms, we introduce Random Oligarchy. The mechanism asks just 3 queries and is essentially optimal among the class of such mechanisms with respect to Distortion. In summary, we demonstrate the surprising power of constant sample complexity mechanisms generally, and just three random voters in particular, to provide some of the best known results in the implicit utilitarian framework.