Magnetic Preference Optimization: Achieving Last-iterate Convergence for Language Models Alignment

Self-play methods have demonstrated remarkable success in enhancing model capabilities across various domains. In the context of Reinforcement Learning from Human Feedback (RLHF), self-play not only boosts Large Language Model (LLM) performance but also overcomes the limitations of traditional Bradley-Terry (BT) model assumptions by finding the Nash equilibrium (NE) of a preference-based, two-player constant-sum game. However, existing methods either guarantee only average-iterate convergence, incurring high storage and inference costs, or converge to the NE of a regularized game, failing to accurately reflect true human preferences. In this paper, we introduce Magnetic Preference Optimization (MPO), a novel approach capable of achieving last-iterate convergence to the NE of the original game, effectively overcoming the limitations of existing methods. Building upon Magnetic Mirror Descent (MMD), MPO attains a linear convergence rate, making it particularly suitable for fine-tuning LLMs. To ensure our algorithm is both theoretically sound and practically viable, we present a simple yet effective implementation that adapts the theoretical insights to the RLHF setting. Empirical results demonstrate that MPO can significantly enhance the performance of LLMs, highlighting the potential of self-play methods in alignment.

Efficient Model-agnostic Alignment via Bayesian Persuasion

With recent advancements in large language models (LLMs), alignment has emerged as an effective technique for keeping LLMs consensus with human intent. Current methods primarily involve direct training through Supervised Fine-tuning (SFT) or Reinforcement Learning from Human Feedback (RLHF), both of which require substantial computational resources and extensive ground truth data. This paper explores an efficient method for aligning black-box large models using smaller models, introducing a model-agnostic and lightweight Bayesian Persuasion Alignment framework. We formalize this problem as an optimization of the signaling strategy from the small model's perspective. In the persuasion process, the small model (Advisor) observes the information item (i.e., state) and persuades large models (Receiver) to elicit improved responses. The Receiver then generates a response based on the input, the signal from the Advisor, and its updated belief about the information item. Through training using our framework, we demonstrate that the Advisor can significantly enhance the performance of various Receivers across a range of tasks. We theoretically analyze our persuasion framework and provide an upper bound on the Advisor's regret, confirming its effectiveness in learning the optimal signaling strategy. Our Empirical results demonstrates that GPT-2 can significantly improve the performance of various models, achieving an average enhancement of 16.1% in mathematical reasoning ability and 13.7% in code generation. We hope our work can provide an initial step toward rethinking the alignment framework from the Bayesian Persuasion perspective.

Foundational Challenges in Assuring Alignment and Safety of Large Language Models

This work identifies 18 foundational challenges in assuring the alignment and safety of large language models (LLMs). These challenges are organized into three different categories: scientific understanding of LLMs, development and deployment methods, and sociotechnical challenges. Based on the identified challenges, we pose $200+$ concrete research questions.

Incentive Compatibility for AI Alignment in Sociotechnical Systems: Positions and Prospects

The burgeoning integration of artificial intelligence (AI) into human society brings forth significant implications for societal governance and safety. While considerable strides have been made in addressing AI alignment challenges, existing methodologies primarily focus on technical facets, often neglecting the intricate sociotechnical nature of AI systems, which can lead to a misalignment between the development and deployment contexts. To this end, we posit a new problem worth exploring: Incentive Compatibility Sociotechnical Alignment Problem (ICSAP). We hope this can call for more researchers to explore how to leverage the principles of Incentive Compatibility (IC) from game theory to bridge the gap between technical and societal components to maintain AI consensus with human societies in different contexts. We further discuss three classical game problems for achieving IC: mechanism design, contract theory, and Bayesian persuasion, in addressing the perspectives, potentials, and challenges of solving ICSAP, and provide preliminary implementation conceptions.

CivRealm: A Learning and Reasoning Odyssey in Civilization for Decision-Making Agents

The generalization of decision-making agents encompasses two fundamental elements: learning from past experiences and reasoning in novel contexts. However, the predominant emphasis in most interactive environments is on learning, often at the expense of complexity in reasoning. In this paper, we introduce CivRealm, an environment inspired by the Civilization game. Civilization's profound alignment with human history and society necessitates sophisticated learning, while its ever-changing situations demand strong reasoning to generalize. Particularly, CivRealm sets up an imperfect-information general-sum game with a changing number of players; it presents a plethora of complex features, challenging the agent to deal with open-ended stochastic environments that require diplomacy and negotiation skills. Within CivRealm, we provide interfaces for two typical agent types: tensor-based agents that focus on learning, and language-based agents that emphasize reasoning. To catalyze further research, we present initial results for both paradigms. The canonical RL-based agents exhibit reasonable performance in mini-games, whereas both RL- and LLM-based agents struggle to make substantial progress in the full game. Overall, CivRealm stands as a unique learning and reasoning challenge for decision-making agents. The code is available at https://github.com/bigai-ai/civrealm.

AI Alignment: A Comprehensive Survey

AI alignment aims to make AI systems behave in line with human intentions and values. As AI systems grow more capable, the potential large-scale risks associated with misaligned AI systems become salient. Hundreds of AI experts and public figures have expressed concerns about AI risks, arguing that "mitigating the risk of extinction from AI should be a global priority, alongside other societal-scale risks such as pandemics and nuclear war". To provide a comprehensive and up-to-date overview of the alignment field, in this survey paper, we delve into the core concepts, methodology, and practice of alignment. We identify the RICE principles as the key objectives of AI alignment: Robustness, Interpretability, Controllability, and Ethicality. Guided by these four principles, we outline the landscape of current alignment research and decompose them into two key components: forward alignment and backward alignment. The former aims to make AI systems aligned via alignment training, while the latter aims to gain evidence about the systems' alignment and govern them appropriately to avoid exacerbating misalignment risks. Forward alignment and backward alignment form a recurrent process where the alignment of AI systems from the forward process is verified in the backward process, meanwhile providing updated objectives for forward alignment in the next round. On forward alignment, we discuss learning from feedback and learning under distribution shift. On backward alignment, we discuss assurance techniques and governance practices that apply to every stage of AI systems' lifecycle. We also release and continually update the website (www.alignmentsurvey.com) which features tutorials, collections of papers, blog posts, and other resources.

Measuring Value Understanding in Language Models through Discriminator-Critique Gap

Recent advancements in Large Language Models (LLMs) have heightened concerns about their potential misalignment with human values. However, evaluating their grasp of these values is complex due to their intricate and adaptable nature. We argue that truly understanding values in LLMs requires considering both "know what" and "know why". To this end, we present the Value Understanding Measurement (VUM) framework that quantitatively assesses both "know what" and "know why" by measuring the discriminator-critique gap related to human values. Using the Schwartz Value Survey, we specify our evaluation values and develop a thousand-level dialogue dataset with GPT-4. Our assessment looks at both the value alignment of LLM's outputs compared to baseline answers and how LLM responses align with reasons for value recognition versus GPT-4's annotations. We evaluate five representative LLMs and provide strong evidence that the scaling law significantly impacts "know what" but not much on "know why", which has consistently maintained a high level. This may further suggest that LLMs might craft plausible explanations based on the provided context without truly understanding their inherent value, indicating potential risks.

ProAgent: Building Proactive Cooperative AI with Large Language Models

Building AIs with adaptive behaviors in human-AI cooperation stands as a pivotal focus in AGI research. Current methods for developing cooperative agents predominantly rely on learning-based methods, where policy generalization heavily hinges on past interactions with specific teammates. These approaches constrain the agent's capacity to recalibrate its strategy when confronted with novel teammates. We propose \textbf{ProAgent}, a novel framework that harnesses large language models (LLMs) to fashion a \textit{pro}active \textit{agent} empowered with the ability to anticipate teammates' forthcoming decisions and formulate enhanced plans for itself. ProAgent excels at cooperative reasoning with the capacity to dynamically adapt its behavior to enhance collaborative efforts with teammates. Moreover, the ProAgent framework exhibits a high degree of modularity and interpretability, facilitating seamless integration to address a wide array of coordination scenarios. Experimental evaluations conducted within the framework of \textit{Overcook-AI} unveil the remarkable performance superiority of ProAgent, outperforming five methods based on self-play and population-based training in cooperation with AI agents. Further, when cooperating with human proxy models, its performance exhibits an average improvement exceeding 10\% compared to the current state-of-the-art, COLE. The advancement was consistently observed across diverse scenarios involving interactions with both AI agents of varying characteristics and human counterparts. These findings inspire future research for human-robot collaborations. For a hands-on demonstration, please visit \url{https://pku-proagent.github.io}.

Heterogeneous Value Evaluation for Large Language Models

The emergent capabilities of Large Language Models (LLMs) have made it crucial to align their values with those of humans. Current methodologies typically attempt alignment with a homogeneous human value and requires human verification, yet lack consensus on the desired aspect and depth of alignment and resulting human biases. In this paper, we propose A2EHV, an Automated Alignment Evaluation with a Heterogeneous Value system that (1) is automated to minimize individual human biases, and (2) allows assessments against various target values to foster heterogeneous agents. Our approach pivots on the concept of value rationality, which represents the ability for agents to execute behaviors that satisfy a target value the most. The quantification of value rationality is facilitated by the Social Value Orientation framework from social psychology, which partitions the value space into four categories to assess social preferences from agents' behaviors. We evaluate the value rationality of eight mainstream LLMs and observe that large models are more inclined to align neutral values compared to those with strong personal values. By examining the behavior of these LLMs, we contribute to a deeper understanding of value alignment within a heterogeneous value system.

STAS: Spatial-Temporal Return Decomposition for Multi-agent Reinforcement Learning

Centralized Training with Decentralized Execution (CTDE) has been proven to be an effective paradigm in cooperative multi-agent reinforcement learning (MARL). One of the major challenges is yet credit assignment, which aims to credit agents by their contributions. Prior studies focus on either implicitly decomposing the joint value function or explicitly computing the payoff distribution of all agents. However, in episodic reinforcement learning settings where global rewards can only be revealed at the end of the episode, existing methods usually fail to work. They lack the functionality of modeling complicated relations of the delayed global reward in the temporal dimension and suffer from large variance and bias. We propose a novel method named Spatial-Temporal Attention with Shapley (STAS) for return decomposition; STAS learns credit assignment in both the temporal and the spatial dimension. It first decomposes the global return back to each time step, then utilizes Shapley Value to redistribute the individual payoff from the decomposed global reward. To mitigate the computational complexity of Shapley Value, we introduce an approximation of marginal contribution and utilize Monte Carlo sampling to estimate Shapley Value. We evaluate our method on the classical Alice & Bob example and Multi-agent Particle Environments benchmarks across different scenarios, and we show our methods achieve an effective spatial-temporal credit assignment and outperform all state-of-art baselines.