Solves reinforcement learning reward design for user-aligned agents.
problem Designing effective reward functions for reinforcement learning.
method Reward modeling: learn from user interaction and optimize.
result Agents can be aligned with user intentions through learned reward functions.
OOM-RL uses financial market losses to align AI agents in autonomous systems.
problem Constrained alignment of autonomous software agents in live financial markets.
method Deploying agents in live financial markets to enforce strict test-driven workflows.
result Final OOM-RL-aligned system achieved a stable equilibrium with an annualized Sharpe ratio of 2.06.
Study examines if LLMs' trading styles match real market behavior.
problem Lack of behavioral consistency in LLMs' trading strategies.
method Year-long simulations with LLMs, operationalizing behavioral finance drivers, and comparing with financial theory.
result LLMs' strategy switching is only partially consistent with behavioral finance theories.
Paper presents a method to efficiently learn ordered representations of multi-agent data.
problem Challenges in learning consistent representations of multi-agent interactions.
method Dynamic alignment method to order multi-agent data for faster representation learning.
result Representation learning of multi-agent data is significantly accelerated.
LLM trading agents show risk feedback can improve alignment without fine-tuning.
problem Aligning LLM trading agents with financial risk.
method TradeArena testbed, risk reports, execution simulation, memory replay.
result Risk feedback can improve alignment without fine-tuning, but not universally.
HawkesLLM models text generation with temporal influence, improving semantic alignment under limited memory.
problem Path-dependent uncertainty in agentic text-simulation systems.
method HawkesLLM framework separates temporal influence modeling from text generation, using a multivariate Hawkes process and a language model.
result HawkesLLM improves late-stage semantic alignment under a compact prompt-memory budget.
GWIL uses Gromov-Wasserstein distance to align expert and imitation agent states.
problem Cross-domain imitation learning challenges due to different system dimensions and stationary distributions.
method Gromov-Wasserstein Imitation Learning (GWIL) using Gromov-Wasserstein distance.
result GWIL effectively aligns expert and imitation agent states in various continuous control domains.
Incentive-aware recommender system for online platforms.
problem Myopic agents exploit optimal arms, not exploring alternatives.
method Model as multi-agent bandit problem, incentivizes exploration.
result Asymptotically optimal performance with ex-post fairness.
New approach categorizes objective functions for embodied agents.
problem Understanding how objectives relate to each other and discovering new objectives.
method Introducing Action Perception Divergence (APD) to categorize objective functions.
result Introduces a spectrum of objectives from narrow to general, explaining various unsupervised objectives.
A system for supervising decentralized finance risks using LLMs and structured evidence.
problem Supervising decentralized finance risks
method Forecast-grounded agentic supervision system
result Developed a system that scores tickets against a regulator-aligned ground truth and false-intervention rate.
Agents learn and control complex mechanical systems through shared memories.
problem Controlling multi-joint dynamical systems.
method Coupled autoregressive active inference agents using Bayesian filtering and minimizing expected free energy.
result Demonstrated learning and control of a double mass-spring-damper system.
Study replicates market model, finds replication hindered by missing details.
problem Replicating a market model with missing details and limited quantitative reporting.
method Increased simulation runs, bootstrap confidence intervals, and code analysis.
result Achieved relational equivalence for most metrics but rejected quantitative alignment.
DeXposure-Claw supervises decentralized finance risks by grounding LLM decisions in evidence.
problem Weak evidence leads to over-interventions by general-purpose LLM agents in decentralized finance.
method DeXposure-Claw uses a graph time-series foundation model to forecast exposure networks, turning forecasts into alerts and constraining escalation with data-health gates.
result DeXposure-Claw reduces false alarms and improves regulator alignment in decentralized finance risk supervision.
The paper develops an economic foundation for multi-agent learning in markets.
problem Learning dynamics in markets with strategic externalities.
method A two-phase incentive mechanism that estimates and uses implementable transfers to steer long-run dynamics.
result The mechanism achieves sublinear social-welfare regret and asymptotically optimal welfare under mild rationality and exploration conditions.
PPO algorithm converges to global optimality in multi-agent reinforcement learning.
problem Designing statistical guarantees for policy optimization methods in multi-agent reinforcement learning.
method Leveraging a multi-agent performance difference lemma, a localized action value function is used as a descent direction for each local policy, leading to a multi-agent PPO algorithm.
result The multi-agent PPO algorithm converges to the globally optimal policy at a sublinear rate under standard regularity conditions.
ADR helps LLMs find and use historical analogies for foresight analysis.
problem LLMs struggle to find relevant historical analogies due to surface-level matching.
method Proposes CANA framework with mechanism alignment and cross-analogy confirmation.
result CANA improves historical analogy generation by up to 10%.
MaxMax Q-Learning improves coordination in multi-agent reinforcement learning by refining action selection.
problem Relative over-generalization in decentralized multi-agent reinforcement learning.
method MaxMax Q-Learning employs iterative sampling and evaluation of potential next states to refine approximations of ideal state transitions.
result MaxMax Q-Learning frequently outperforms existing baselines, demonstrating enhanced convergence and sample efficiency.
FedSight AI predicts federal funds rate using LLMs and multi-agent reasoning.
problem Predicting Federal Open Market Committee's decisions on federal funds rate.
method Multi-agent framework with large language models, structured and unstructured inputs, and CoD extension for efficient reasoning.
result Achieved 93.75% accuracy and 93.33% stability in predicting FOMC outcomes.
SEMI uses multisensory incongruity to self-supervise exploration in reinforcement learning.
problem Efficient exploration in reinforcement learning with sparse or missing rewards.
method SEMI incentivizes exploration by maximizing multisensory incongruity, measured in perception and action incongruity.
result SEMI improves sample efficiency and learns skills without external rewards.
This study analyzes mutual influence on investment strategies of financial market agents.
problem Mutual influence among agents in financial markets and its impact on investment strategies.
method Formulated optimal investment differential game problem, derived analytical solutions, proposed fast algorithm, and theoretically analyzed mutual influence.
result Agents' optimal strategies converge to the asymptotic strategy when mutual influence is strong and approaches infinity.
Study shows market volatility affects optimal communication design for trading strategies.
problem Investigating how communication impacts trading strategy performance in multi-agent systems.
method 5-agent LLM-based trading systems across 450 experiments spanning 21 months, comparing 5 organizational structures.
result Communication improves performance but depends on market characteristics, with competitive conversation excelling in volatile tech stocks.
A test measures artificial agents' human-like behavior in video games.
problem Measuring the believability of artificial agents' human-like behavior.
method Developed a non-parametric two-sample hypothesis test.
result The p p p -value correlates with human judgment of human-like behavior. This study proposes hidden state curiosity to enhance RL models' resilience against noise.
problem Curiosity traps in RL models distract agents from discovering novel experiences.
method Proposed hidden state curiosity based on the Free Energy Principle to reward agents for KL divergence between predictive priors and posteriors.
result Agents with hidden state curiosity are more resilient against curiosity traps compared to those with prediction error curiosity.
Study proposes a multi-agent system using LLMs for REIT trading, outperforming benchmarks.
problem Low-volatility Chinese REIT market, low risk-adjusted returns.
method Multi-agent framework with four types of agents, prediction model pathways, fine-tuning.
result Multi-agent strategies outperform buy-and-hold in terms of return, Sharpe ratio, and drawdown.
ARCO-BO optimizes multi-agent design under heterogeneity, improving efficiency and performance.
problem Heterogeneous multi-agent optimization challenges in resource use and information sharing.
method ARCO-BO integrates a consensus mechanism, budget-aware sampling, and partial input sharing for heterogeneous design spaces.
result ARCO-BO outperforms independent and collaborative BO methods in complex multi-agent settings.
RD-Agent(Q) automates quantitative finance research and development.
problem Challenges in asset return prediction due to high dimensionality and volatility.
method Data-centric multi-agent framework for automated research and development of quantitative strategies.
result Up to 2X higher annualized returns with 70% fewer factors.
Robust RL with learned optimal adversary improves agent performance under adversarial state observations.
problem Ensuring reinforcement learning agents' robustness against adversarial perturbations of state observations.
method Proposed a framework of alternating training with learned adversaries (ATLA) to find optimal adversarial policies and enhance agent robustness.
result ATLA achieves state-of-the-art performance under strong adversaries in continuous control environments.
Improved investment performance with fine-grained LLM tasks.
problem Abstract financial trading systems often overlook real-world workflow intricacies, leading to degraded performance.
method Proposes a multi-agent LLM trading framework that decomposes investment analysis into fine-grained tasks.
result Fine-grained task decomposition significantly improves risk-adjusted returns compared to coarse-grained designs.
KVCOMM optimizes multi-agent LLM systems by reusing KV-caches, reducing redundant processing.
problem Substantial overhead from reprocessing overlapping contexts across multi-agent systems.
method KVCOMM reuses KV-caches and adjusts offsets for shared content using a pool of cached examples (anchors).
result Achieves over 70% reuse rate across diverse multi-agent tasks, up to 7.8x speedup.
This work tackles robust RL in multi-agent settings, improving sample efficiency.
problem Overcoming environmental uncertainties in multi-agent reinforcement learning.
method Proposes DRNVI, a sample-efficient algorithm for learning robust equilibria in RMGs.
result Establishes near-optimal sample complexity for solving RMGs.
Paper reviews the evolution of alpha from human insight to AI-powered systems.
problem Exceeding market benchmarks in finance.
method Five-stage taxonomy integrating representation learning, multimodal data fusion, and LLM agents.
result Unified framework for evaluating and developing next-gen alpha systems.
FinMem enhances LLM trading agents with layered memory and character design.
problem Developing purpose-driven LLM agents for financial decision-making.
method Integrates layered memory and character design modules into an LLM framework.
result Significantly enhanced trading performance in financial markets.
Collective motion is an intriguing phenomenon, especially considering that it arises from a set of simple rules governing local interactions between individuals. In theoretical models, these rules are normally \emph{assumed} to take a particular form, possibly constrained by heuristic arguments. We propose a new class …
Study uses LLMs for financial sentiment analysis without fine-tuning.
problem Lack of prescriptive knowledge to leverage generative models in FSA.
method Proposes a design framework with heterogeneous LLM agents based on Minsky's theory.
result Framework yields better accuracies, especially with substantial discussions.
Agents trained with reinforcement learning deviate from Nash equilibrium in optimal execution game.
problem Deviation of reinforcement learning strategies from Nash equilibrium in optimal execution game.
method Two-player optimal execution game with reinforcement learning algorithms (Double Deep Q-Learning).
result Strategies learned by agents deviate significantly from Nash equilibrium, exhibiting supra-competitive solutions.
New algorithm learns coordinated decisions in loosely-coupled multi-agent systems.
problem Learning coordinated decisions in multi-agent systems with sparse interactions.
method Multi-Agent Thompson Sampling (MATS) for multi-agent multi-armed bandits.
result MATS achieves sublinear regret and outperforms MAUCE on synthetic and real benchmarks.
Enhanced financial trading system using multi-agent LLMs with layered memory.
problem Inefficient prioritization of tasks in LLMs due to their memory processing.
method Introducing a multi-agent framework with layered memories and inter-agent debate.
result Superior automated trading accuracy and decision robustness.
Study on RL on volatility surfaces, proving no free lunch for law-seeking methods.
problem Aligning RL agents with no-arbitrage laws in volatile markets.
method Built a law manifold, defined penalties, and used a Goodhart decomposition.
result No free lunch theorem: Law-seeking RL cannot outperform baselines.
Dreamer 4 learns Minecraft tasks from videos alone.
problem Accurately predicting object interactions in complex environments.
method Reinforcement learning inside a fast, accurate world model.
result Dreamer 4 outperforms previous models in Minecraft, learning from only offline data.
Study allocates resources to strategic agents while balancing cost and incentives.
problem Dynamic allocation of reusable resources to strategic agents with private valuations under long-term cost constraints.
method Incentive-aware framework combining epoch-based lazy updates and randomized exploration rounds.
result Achieves i l d e O ( T ) ilde{\mathcal{O}}(\sqrt{T}) i l d e O ( T ) social welfare regret, satisfies all cost constraints, and ensures incentive alignment. GeomHerd predicts herding behavior before market prices move, using Ricci curvature of agent interaction graphs.
problem Quantifying herding behavior in markets that lags behind actual price movements.
method Develops a geometric framework to track coordination on agent interaction graphs, bypassing lag in price-correlation statistics.
result GeomHerd anticipates herding long before market baselines, with significant lead times in predictions.
To make AI systems broadly useful for challenging real-world tasks, we need them to learn complex human goals and preferences. One approach to specifying complex goals asks humans to judge during training which agent behaviors are safe and useful, but this approach can fail if the task is too complicated for a human to…
A new method for optimizing language-based agentic systems using semantic backpropagation.
problem Lack of proper feedback assignment in optimizing agentic systems.
method Formalization of semantic backpropagation with semantic gradients and semantic gradient descent.
result Our method outperforms existing state-of-the-art methods for solving GASO problems.
Paper introduces a novel reward function for noisy financial markets using imitation learning.
problem Noisy reward function in financial markets hinders RL agent performance.
method Integrates imitation learning feedback with reinforcement learning to improve reward function design.
result Improves financial performance metrics compared to traditional benchmarks and RL agents.
Core-Halo solves large-scale fixed-point problems by decentralizing updates.
problem Large-scale fixed-point equations with block dependencies.
method Core-Halo decomposition separates write ownership from read-only context, aligning with block-dependence structure.
result Core-Halo achieves near-centralized performance while retaining parallelism.
New approach treats coordination as an architectural layer to improve LLM-based multi-agent systems.
problem Coordination defects lead to high failure rates in LLM-based multi-agent systems.
method Treats coordination as a configurable architectural layer separable from agent logic and information access.
result Configurations leave distinguishable signatures, enabling architectural reasoning and Pareto frontiers.
A novel approach learns constraints and maximizes rewards for autonomous agents.
problem Ensuring autonomous agents align with societal norms and values.
method Inverse reinforcement learning for constraints, contextual bandit orchestrator for policy mixing.
result Agent learns to act optimally within constraints and maximize rewards.
Calibrated models can lead to miscalibrated aggregations in strategic interactions.
problem Miscalibration in aggregated predictions from multiple calibrated models.
method Analysis of strategic interactions between calibrated predictors, proving conditions for miscalibration and comparing VCG and Brier-score aggregation methods.
result VCG aggregation method outperforms Brier-score in strategic settings, providing robustness and comparable accuracy.