Reward hacking exploits misspecified rewards, affecting agent capabilities and true performance.
problem Reward hacking in RL models exploiting reward misspecifications.
method Constructed four RL environments with misspecified rewards; analyzed agent capabilities and behavior.
result More capable agents exploit reward misspecifications, achieving higher proxy reward but lower true reward.
Agents need world models to generalize multi-step tasks.
problem The necessity of world models for flexible, goal-directed behavior.
method Formal analysis and demonstration of the necessity of world models for agents to generalize multi-step tasks.
result World models are necessary for agents to generalize to multi-step goal-directed tasks.
This paper explores how LLMs can improve pipeline-based conversational agents.
problem Limitations of pipeline-based conversational agents in human-like conversations.
method Investigated LLMs' capabilities in two phases: design and development, and operations.
result LLMs can enhance pipeline-based agents in various tasks like data generation, intent classification, and auto-correction.
For autonomous vehicles (AVs) to behave appropriately on roads populated by human-driven vehicles, they must be able to reason about the uncertain intentions and decisions of other drivers from rich perceptual information. Towards these capabilities, we present a probabilistic forecasting model of future interactions b…
Framework analyzes RL agents' behavior to explain their strengths and weaknesses.
problem Understanding and explaining RL agents' capabilities and limitations.
method Data analysis and visualization of interaction history to extract interestingness elements.
result Visual summaries help humans correctly perceive agents' strengths and weaknesses.
Transformers learn to play games in-context, proving Nash equilibrium.
problem Understanding in-context game-playing capabilities of pre-trained transformers.
method Theoretical guarantees and constructional results for transformer architecture in multi-agent games.
result Pre-trained transformers can learn Nash equilibrium in-context for two-player zero-sum games.
We present an agent based model of a single asset financial market that is capable of replicating several non-trivial statistical properties observed in real financial markets, generically referred to as stylized facts. While previous models reported in the literature are also capable of replicating some of these stati…
CausalGame benchmarks LLM agents' causal thinking in games.
problem Evaluating causal thinking in AI Scientists with LLMs.
method Interactive games with 14 scenarios incorporating selection bias, measurement error, and hidden confounders.
result None of the 30 LLM agents demonstrated reliable causal thinking, with the best model achieving only 68.0% survival.
The security of Deep Reinforcement Learning (Deep RL) algorithms deployed in real life applications are of a primary concern. In particular, the robustness of RL agents in cyber-physical systems against adversarial attacks are especially vital since the cost of a malevolent intrusions can be extremely high. Studies hav…
New modeling approach for self-organizing complex systems.
problem Identifying general principles of self-organizing complex systems.
method Self-deploying system structure and activities for goal-driven agents.
result Self-organization emerges from rational activity algorithm based on goals dependency network.
Scaling up model and data size improves imitation learning in single-agent games.
problem Limited recovery of expert behavior in single-agent games using imitation learning.
method Investigate the effect of scaling model and data size on imitation learning performance.
result IL loss and mean return scale with compute budget, resulting in power laws.
A multi-agent system improves crypto portfolio management by processing diverse data types.
problem Managing cryptocurrency portfolios requires processing various data types under high volatility.
method A multi-agent system with three specialized agents for market dynamics, news sentiment, and signal fusion.
result The best configuration, Hierarchical (Skill), achieved a 133.52% cumulative return and 1.502 Sharpe ratio.
We present an effective technique for training deep learning agents capable of negotiating on a set of clauses in a contract agreement using a simple communication protocol. We use Multi Agent Reinforcement Learning to train both agents simultaneously as they negotiate with each other in the training environment. We al…
Agents collaboratively learn optimal policies in MDPs with limited capabilities.
problem Learning optimal policies in MDPs with heterogeneous agents and limited communication.
method Introduced concepts of leakage probabilities and proposed Federated-Q protocol (FedQ) for collaborative learning.
result FedQ protocol effectively aggregates knowledge and modifies learning problems for further training.
Transformer models can approximate smooth functions with prompts, enhancing LLMs' dynamic capabilities.
problem Lack of theoretical framework for prompt engineering in transformer models.
method Formal framework demonstrating transformer models can approximate β-times differentiable functions with prompts. result Transformer models can approximate β-times differentiable functions with arbitrary precision using appropriately structured prompts. FinVision uses LLM agents to predict stock markets by processing various financial data types.
problem Challenges in integrating diverse financial data for accurate stock market prediction.
method Multi-agent framework with LLMs specialized in different financial data types and a reflection module.
result The reflection module enhances decision-making capabilities for financial trading.
We propose a new model of minority game with so-called smart agents such that the standard deviation and the total loss in this model reach the theoretical minimum values in the limit of long time. The smart agents use trail and error method to make a choice but bring global optimization to the system, which suggests t…
Study uses AI agents to improve equity portfolio management.
problem Improving stock selection and portfolio management efficiency.
method Role-based multi-agent systems for equity research.
result Multi-agent approach outperforms benchmarks in stock selection.
We investigate a classification problem using multiple mobile agents capable of collecting (partial) pose-dependent observations of an unknown environment. The objective is to classify an image over a finite time horizon. We propose a network architecture on how agents should form a local belief, take local actions, an…
Policy-gradient method controls multiple non-cohesive targets.
problem Controlling multiple non-cohesive targets in a decentralized manner.
method Proximal Policy Optimization for target selection and driving.
result Effective control of non-cohesive targets without prior dynamics knowledge.
YRC-Bench benchmarks AI agents learning to collaborate with experts.
problem Learning to recognize when to consult an expert in novel situations.
method Validation strategy and proposer-validator decomposition.
result Insights for future AI safety research.
Economies and societal structures in general are complex stochastic systems which may not lend themselves well to algebraic analysis. An addition of subjective value criteria to the mechanics of interacting agents will further complicate analysis. The purpose of this short study is to demonstrate capabilities of agent-…
New theorems show agents need specific internal structures to perform well under uncertainty.
problem How do agents need to be structured to perform well under uncertainty?
method Proved selection theorems showing strong task performance forces specific internal structures.
result Strong task performance forces world models, belief-like memory, and persistent regime-tracking variables.
Due to the capability of deep learning to perform well in high dimensional problems, deep reinforcement learning agents perform well in challenging tasks such as Atari 2600 games. However, clearly explaining why a certain action is taken by the agent can be as important as the decision itself. Deep reinforcement learni…
New RL agent learns sparse rewards efficiently.
problem Sparse-reward environments and computational expense.
method Active inference with novel free energy minimization.
result High sample efficiency and online operation.
AI-Trader benchmarks LLMs in live financial markets, revealing poor trading performance.
problem Challenges in real-time financial decision-making by autonomous agents.
method Fully automated, live evaluation benchmark with minimal human intervention.
result General intelligence does not translate to effective trading, highlighting limitations.
This paper tackles sandbagging in AI safety evaluations.
problem AI agents may hide dangerous capabilities to avoid being deactivated.
method Developed a simple model of strategic deception in sequential decision-making tasks.
result Demonstrated that optimal rational agents exhibit sandbagging behavior.
In this paper, we investigate how to learn to control a group of cooperative agents with limited sensing capabilities such as robot swarms. The agents have only very basic sensor capabilities, yet in a group they can accomplish sophisticated tasks, such as distributed assembly or search and rescue tasks. Learning a pol…
Through multi-agent competition, the simple objective of hide-and-seek, and standard reinforcement learning algorithms at scale, we find that agents create a self-supervised autocurriculum inducing multiple distinct rounds of emergent strategy, many of which require sophisticated tool use and coordination. We find clea…
Causal reasoning has been an indispensable capability for humans and other intelligent animals to interact with the physical world. In this work, we propose to endow an artificial agent with the capability of causal reasoning for completing goal-directed tasks. We develop learning-based approaches to inducing causal kn…
The ability to generalize is an important feature of any intelligent agent. Not only because it may allow the agent to cope with large amounts of data, but also because in some environments, an agent with no generalization capabilities cannot learn. In this work we outline several criteria for generalization, and prese…
We present a model in which we investigate the structure and evolution of a random network that connects agents capable of exchanging wealth. Economic interactions between neighbors can occur only if the difference between their wealth is less than a threshold value that defines the width of the economic classes. If th…
A new multi-agent learning method improves performance in complex games.
problem Performance gap between MAPG and value-based multi-agent approaches.
method Introduces value function decomposition into multi-agent actor-critic framework for off-policy learning.
result DOP significantly outperforms state-of-the-art multi-agent reinforcement learning algorithms.
Blockchain helps secure payments between AI agents.
problem Ensuring secure payments between untrusted AI agents.
method Systematized four-stage lifecycle for A2A payments on blockchain.
result Challenges remain in weak intent binding, misuse, and limited accountability.
Survey examines agentic AI in finance, highlighting its autonomy and challenges.
problem Autonomous AI systems in finance and their implications.
method Systematic review of research, technical architectures, market applications, and governance frameworks.
result Agentic AI offers enhanced market efficiency but introduces new risks.
MPLP learns neural network weights by treating operations as message-passing agents.
problem Training neural networks using gradient-based methods.
method MPLP abstracts neural network operations as message-passing agents, updating internal states and passing messages.
result MPLP outperforms traditional gradient-based methods on simple feed-forward neural networks.
PyFi uses adversarial agents to train VLMs on financial image understanding.
problem Training VLMs to understand complex financial questions.
method PyFi-600K dataset and adversarial MCTS mechanism.
result Fine-tuned VLMs improve by 19.52% and 8.06% on financial question accuracy.
Predicts language model performance from public models without training.
problem Understanding how language model performance varies with scale.
method Observational approach building scaling laws from publicly available models.
result Smooth, sigmoidal behavior of emergent phenomena is predictable from small models.
A novel framework for adaptive multi-agent communication in reinforcement learning.
problem Manual specification of communication structures in multi-agent reinforcement learning.
method Learning Structured Communication (LSC) framework using hierarchical graph neural networks.
result Adaptive hierarchical formations and efficient message propagation among agents.
InvestorBench benchmarks LLM agents in financial tasks.
problem Lack of a comprehensive benchmark for LLM-based financial agents.
method Developed a benchmark with diverse financial tasks and datasets.
result Evaluated LLM agents' performance across various financial products and market environments.
In this work we describe a novel deep reinforcement learning architecture that allows multiple actions to be selected at every time-step in an efficient manner. Multi-action policies allow complex behaviours to be learnt that would otherwise be hard to achieve when using single action selection techniques. We use both …
Algorithm for decentralized competition among adaptive agents.
problem Decentralized competition among adaptive networks.
method Developed an algorithm for decentralized competition among teams of adaptive agents.
result Algorithm enables decentralized competition among adaptive agents.
Enhances RL agents with predictive internal representations.
problem Improving model-free reinforcement learning agents.
method Introduces Deep InfoMax (DIM) objective to train predictive internal representations.
result Successfully learned predictive representations in synthetic settings.
This paper investigates generalisation in multi-agent games, where the generality of the agent can be evaluated by playing against opponents it hasn't seen during training. We propose two new games with concealed information and complex, non-transitive reward structure (think rock/paper/scissors). It turns out that mos…
Reinforcement learning (RL) algorithms have been around for decades and employed to solve various sequential decision-making problems. These algorithms however have faced great challenges when dealing with high-dimensional environments. The recent development of deep learning has enabled RL methods to drive optimal pol…
PandaAI: A practical agent for neuro-symbolic data analysis and decision-making in finance
problem Sequential decision-making in finance
method Leveraging LLMs for market regime modeling and constrained alpha generation
result PandaAI achieves higher Rank IC and lower maximum drawdown
New methods improve multi-agent reinforcement learning by addressing rotational dynamics.
problem Reproducibility crisis in multi-agent reinforcement learning.
method Reframing MARL approaches using Variational Inequalities (VIs) and proposing gradient-based VI methods.
result Significant performance improvements across benchmarks, including better convergence to equilibrium strategies in zero-sum games.
The study proves necessary conditions for robust decision-making in uncertain environments.
problem Conditions for robust decision-making in uncertain environments.
method Quantitative selection theorems and binary betting decisions.
result World models, belief-like memory, and persistent variables are necessary for strong task performance.