Royal Holloway, University of London
Abstract:We introduce the Logic-Enhanced Language Model Agents (LELMA) framework, a novel approach to enhance the trustworthiness of social simulations that utilize large language models (LLMs). While LLMs have gained attention as agents for simulating human behaviour, their applicability in this role is limited by issues such as inherent hallucinations and logical inconsistencies. LELMA addresses these challenges by integrating LLMs with symbolic AI, enabling logical verification of the reasoning generated by LLMs. This verification process provides corrective feedback, refining the reasoning output. The framework consists of three main components: an LLM-Reasoner for producing strategic reasoning, an LLM-Translator for mapping natural language reasoning to logic queries, and a Solver for evaluating these queries. This study focuses on decision-making in game-theoretic scenarios as a model of human interaction. Experiments involving the Hawk-Dove game, Prisoner's Dilemma, and Stag Hunt highlight the limitations of state-of-the-art LLMs, GPT-4 Omni and Gemini 1.0 Pro, in producing correct reasoning in these contexts. LELMA demonstrates high accuracy in error detection and improves the reasoning correctness of LLMs via self-refinement, particularly in GPT-4 Omni.
Abstract:This paper bridges the gap between mathematical heuristic strategies learned from Deep Reinforcement Learning (DRL) in automated agent negotiation, and comprehensible, natural language explanations. Our aim is to make these strategies more accessible to non-experts. By leveraging traditional Natural Language Processing (NLP) techniques and Large Language Models (LLMs) equipped with Transformers, we outline how parts of DRL strategies composed of parts within strategy templates can be transformed into user-friendly, human-like English narratives. To achieve this, we present a top-level algorithm that involves parsing mathematical expressions of strategy templates, semantically interpreting variables and structures, generating rule-based primary explanations, and utilizing a Generative Pre-trained Transformer (GPT) model to refine and contextualize these explanations. Subsequent customization for varied audiences and meticulous validation processes in an example illustrate the applicability and potential of this approach.
Abstract:The evolution of Generative Pre-trained Transformer (GPT) models has led to significant advancements in various natural language processing applications, particularly in legal textual entailment. We present an analysis of GPT-3.5 (ChatGPT) and GPT-4 performances on COLIEE Task 4 dataset, a prominent benchmark in this domain. The study encompasses data from Heisei 18 (2006) to Reiwa 3 (2021), exploring the models' abilities to discern entailment relationships within Japanese statute law across different periods. Our preliminary experimental results unveil intriguing insights into the models' strengths and weaknesses in handling legal textual entailment tasks, as well as the patterns observed in model performance. In the context of proprietary models with undisclosed architectures and weights, black-box analysis becomes crucial for evaluating their capabilities. We discuss the influence of training data distribution and the implications on the models' generalizability. This analysis serves as a foundation for future research, aiming to optimize GPT-based models and enable their successful adoption in legal information extraction and entailment applications.
Abstract:Negation is a fundamental aspect of natural language, playing a critical role in communication and comprehension. Our study assesses the negation detection performance of Generative Pre-trained Transformer (GPT) models, specifically GPT-2, GPT-3, GPT-3.5, and GPT-4. We focus on the identification of negation in natural language using a zero-shot prediction approach applied to our custom xNot360 dataset. Our approach examines sentence pairs labeled to indicate whether the second sentence negates the first. Our findings expose a considerable performance disparity among the GPT models, with GPT-4 surpassing its counterparts and GPT-3.5 displaying a marked performance reduction. The overall proficiency of the GPT models in negation detection remains relatively modest, indicating that this task pushes the boundaries of their natural language understanding capabilities. We not only highlight the constraints of GPT models in handling negation but also emphasize the importance of logical reliability in high-stakes domains such as healthcare, science, and law.
Abstract:We present World of Bugs (WOB), an open platform that aims to support Automated Bug Detection (ABD) research in video games. We discuss some open problems in ABD and how they relate to the platform's design, arguing that learning-based solutions are required if further progress is to be made. The platform's key feature is a growing collection of common video game bugs that may be used for training and evaluating ABD approaches.
Abstract:Automated Bug Detection (ABD) in video games is composed of two distinct but complementary problems: automated game exploration and bug identification. Automated game exploration has received much recent attention, spurred on by developments in fields such as reinforcement learning. The complementary problem of identifying the bugs present in a player's experience has for the most part relied on the manual specification of rules. Although it is widely recognised that many bugs of interest cannot be identified with such methods, little progress has been made in this direction. In this work we show that it is possible to identify a range of perceptual bugs using learning-based methods by making use of only the rendered game screen as seen by the player. To support our work, we have developed World of Bugs (WOB) an open platform for testing ABD methods in 3D game environments.
Abstract:We present a novel bilateral negotiation model that allows a self-interested agent to learn how to negotiate over multiple issues in the presence of user preference uncertainty. The model relies upon interpretable strategy templates representing the tactics the agent should employ during the negotiation and learns template parameters to maximize the average utility received over multiple negotiations, thus resulting in optimal bid acceptance and generation. Our model also uses deep reinforcement learning to evaluate threshold utility values, for those tactics that require them, thereby deriving optimal utilities for every environment state. To handle user preference uncertainty, the model relies on a stochastic search to find user model that best agrees with a given partial preference profile. Multi-objective optimization and multi-criteria decision-making methods are applied at negotiation time to generate Pareto-optimal outcomes thereby increasing the number of successful (win-win) negotiations. Rigorous experimental evaluations show that the agent employing our model outperforms the winning agents of the 10th Automated Negotiating Agents Competition (ANAC'19) in terms of individual as well as social-welfare utilities.
Abstract:With the aim of designing automated tools that assist in the video game quality assurance process, we frame the problem of identifying bugs in video games as an anomaly detection (AD) problem. We develop State-State Siamese Networks (S3N) as an efficient deep metric learning approach to AD in this context and explore how it may be used as part of an automated testing tool. Finally, we show by empirical evaluation on a series of Atari games, that S3N is able to learn a meaningful embedding, and consequently is able to identify various common types of video game bugs.
Abstract:We present a novel negotiation model that allows an agent to learn how to negotiate during concurrent bilateral negotiations in unknown and dynamic e-markets. The agent uses an actor-critic architecture with model-free reinforcement learning to learn a strategy expressed as a deep neural network. We pre-train the strategy by supervision from synthetic market data, thereby decreasing the exploration time required for learning during negotiation. As a result, we can build automated agents for concurrent negotiations that can adapt to different e-market settings without the need to be pre-programmed. Our experimental evaluation shows that our deep reinforcement learning-based agents outperform two existing well-known negotiation strategies in one-to-many concurrent bilateral negotiations for a range of e-market settings.
Abstract:Gaussian Processes (GPs) are widely used tools in statistics, machine learning, robotics, computer vision, and scientific computation. However, despite their popularity, they can be difficult to apply; all but the simplest classification or regression applications require specification and inference over complex covariance functions that do not admit simple analytical posteriors. This paper shows how to embed Gaussian processes in any higher-order probabilistic programming language, using an idiom based on memoization, and demonstrates its utility by implementing and extending classic and state-of-the-art GP applications. The interface to Gaussian processes, called gpmem, takes an arbitrary real-valued computational process as input and returns a statistical emulator that automatically improve as the original process is invoked and its input-output behavior is recorded. The flexibility of gpmem is illustrated via three applications: (i) robust GP regression with hierarchical hyper-parameter learning, (ii) discovering symbolic expressions from time-series data by fully Bayesian structure learning over kernels generated by a stochastic grammar, and (iii) a bandit formulation of Bayesian optimization with automatic inference and action selection. All applications share a single 50-line Python library and require fewer than 20 lines of probabilistic code each.