In the rapidly advancing field of multi-agent systems, ensuring robustness in unfamiliar and adversarial settings is crucial. Notwithstanding their outstanding performance in familiar environments, these systems often falter in new situations due to overfitting during the training phase. This is especially pronounced in settings where both cooperative and competitive behaviours are present, encapsulating a dual nature of overfitting and generalisation challenges. To address this issue, we present Multi-Agent Diagnostics for Robustness via Illuminated Diversity (MADRID), a novel approach for generating diverse adversarial scenarios that expose strategic vulnerabilities in pre-trained multi-agent policies. Leveraging the concepts from open-ended learning, MADRID navigates the vast space of adversarial settings, employing a target policy's regret to gauge the vulnerabilities of these settings. We evaluate the effectiveness of MADRID on the 11vs11 version of Google Research Football, one of the most complex environments for multi-agent reinforcement learning. Specifically, we employ MADRID for generating a diverse array of adversarial settings for TiZero, the state-of-the-art approach which "masters" the game through 45 days of training on a large-scale distributed infrastructure. We expose key shortcomings in TiZero's tactical decision-making, underlining the crucial importance of rigorous evaluation in multi-agent systems.
相關內容
A new online multiple testing procedure is described in the context of anomaly detection, which controls the False Discovery Rate (FDR). An accurate anomaly detector must control the false positive rate at a prescribed level while keeping the false negative rate as low as possible. However in the online context, such a constraint remains highly challenging due to the usual lack of FDR control: the online framework makes it impossible to use classical multiple testing approaches such as the Benjamini-Hochberg (BH) procedure, which would require knowing the entire time series. The developed strategy relies on exploiting the local control of the ``modified FDR'' (mFDR) criterion. It turns out that the local control of mFDR enables global control of the FDR over the full series up to additional modifications of the multiple testing procedures. An important ingredient in this control is the cardinality of the calibration dataset used to compute the empirical p-values. A dedicated strategy for tuning this parameter is designed for achieving the prescribed FDR control over the entire time series. The good statistical performance of the full strategy is analyzed by theoretical guarantees. Its practical behavior is assessed by several simulation experiments which support our conclusions.
In fact-checking, structure and phrasing of claims critically influence a model's ability to predict verdicts accurately. Social media content in particular rarely serves as optimal input for verification systems, which necessitates pre-processing to extract the claim from noisy context before fact checking. Prior work suggests extracting a claim representation that humans find to be checkworthy and verifiable. This has two limitations: (1) the format may not be optimal for a fact-checking model, and (2), it requires annotated data to learn the extraction task from. We address both issues and propose a method to extract claims that is not reliant on labeled training data. Instead, our self-adaptive approach only requires a black-box fact checking model and a generative language model (LM). Given a tweet, we iteratively optimize the LM to generate a claim paraphrase that increases the performance of a fact checking model. By learning from preference pairs, we align the LM to the fact checker using direct preference optimization. We show that this novel setup extracts a claim paraphrase that is more verifiable than their original social media formulations, and is on par with competitive baselines. For refuted claims, our method consistently outperforms all baselines.
Federated learning is highly susceptible to model poisoning attacks, especially those meticulously crafted for servers. Traditional defense methods mainly focus on updating assessments or robust aggregation against manually crafted myopic attacks. When facing advanced attacks, their defense stability is notably insufficient. Therefore, it is imperative to develop adaptive defenses against such advanced poisoning attacks. We find that benign clients exhibit significantly higher data distribution stability than malicious clients in federated learning in both CV and NLP tasks. Therefore, the malicious clients can be recognized by observing the stability of their data distribution. In this paper, we propose AdaAggRL, an RL-based Adaptive Aggregation method, to defend against sophisticated poisoning attacks. Specifically, we first utilize distribution learning to simulate the clients' data distributions. Then, we use the maximum mean discrepancy (MMD) to calculate the pairwise similarity of the current local model data distribution, its historical data distribution, and global model data distribution. Finally, we use policy learning to adaptively determine the aggregation weights based on the above similarities. Experiments on four real-world datasets demonstrate that the proposed defense model significantly outperforms widely adopted defense models for sophisticated attacks.
Quantum computation leverages the use of quantumly-controlled conditionals in order to achieve computational advantage. However, since the different branches in the conditional may operate on the same qubits, a typical approach to compilation involves performing the branches sequentially, which can easily lead to an exponential blowup of the program complexity. We introduce and study a compilation technique for avoiding branch sequentialization in a language that is sound and complete for quantum polynomial time, improving on previously existing polynomialsize bounds and showing the existence of techniques that preserve the intuitive complexity of the program.
We propose a fully Bayesian approach to wideband, or broadband, direction-of-arrival (DoA) estimation and signal detection. Unlike previous works in wideband DoA estimation and detection, where the signals were modeled in the time-frequency domain, we directly model the time-domain representation and treat the non-causal part of the source signal as latent variables. Furthermore, our Bayesian model allows for closed-form marginalization of the latent source signals by leveraging conjugacy. To further speed up computation, we exploit the sparse ``stripe matrix structure'' of the considered system, which stems from the circulant matrix representation of linear time-invariant (LTI) systems. This drastically reduces the time complexity of computing the likelihood from $\mathcal{O}(N^3 k^3)$ to $\mathcal{O}(N k^3)$, where $N$ is the number of samples received by the array and $k$ is the number of sources. These computational improvements allow for efficient posterior inference through reversible jump Markov chain Monte Carlo (RJMCMC). We use the non-reversible extension of RJMCMC (NRJMCMC), which often achieves lower autocorrelation and faster convergence than the conventional reversible variant. Detection, estimation, and reconstruction of the latent source signals can then all be performed in a fully Bayesian manner through the samples drawn using NRJMCMC. We evaluate the detection performance of the procedure by comparing against generalized likelihood ratio testing (GLRT) and information criteria.
Understanding treatment effect heterogeneity is crucial for reliable decision-making in treatment evaluation and selection. The conditional average treatment effect (CATE) is widely used to capture treatment effect heterogeneity induced by observed covariates and to design individualized treatment policies. However, it is an average metric within subpopulations, which prevents it from revealing individual-level risks, potentially leading to misleading results. This article fills this gap by examining individual risk for binary outcomes, specifically focusing on the fraction negatively affected (FNA), a metric that quantifies the percentage of individuals experiencing worse outcomes under treatment compared with control. Even under the strong ignorability assumption, FNA is still unidentifiable, and the existing Frechet-Hoeffding bounds are usually too wide and attainable only under extreme data-generating processes. By invoking mild conditions on the value range of the Pearson correlation coefficient between potential outcomes, we obtain improved bounds compared with previous studies. We show that paradoxically, even with a positive CATE, the lower bound on FNA can be positive, i.e., in the best-case scenario many units will be harmed if they receive treatment. Additionally, we establish a nonparametric sensitivity analysis framework for FNA using the Pearson correlation coefficient as the sensitivity parameter, thereby exploring the relationships among the correlation coefficient, FNA, and CATE. We also propose a method for selecting the range of correlation coefficients. Furthermore, we propose nonparametric estimators for the refined FNA bounds and prove their consistency and asymptotic normality.
Recent artificial intelligence (AI) systems have reached milestones in "grand challenges" ranging from Go to protein-folding. The capability to retrieve medical knowledge, reason over it, and answer medical questions comparably to physicians has long been viewed as one such grand challenge. Large language models (LLMs) have catalyzed significant progress in medical question answering; Med-PaLM was the first model to exceed a "passing" score in US Medical Licensing Examination (USMLE) style questions with a score of 67.2% on the MedQA dataset. However, this and other prior work suggested significant room for improvement, especially when models' answers were compared to clinicians' answers. Here we present Med-PaLM 2, which bridges these gaps by leveraging a combination of base LLM improvements (PaLM 2), medical domain finetuning, and prompting strategies including a novel ensemble refinement approach. Med-PaLM 2 scored up to 86.5% on the MedQA dataset, improving upon Med-PaLM by over 19% and setting a new state-of-the-art. We also observed performance approaching or exceeding state-of-the-art across MedMCQA, PubMedQA, and MMLU clinical topics datasets. We performed detailed human evaluations on long-form questions along multiple axes relevant to clinical applications. In pairwise comparative ranking of 1066 consumer medical questions, physicians preferred Med-PaLM 2 answers to those produced by physicians on eight of nine axes pertaining to clinical utility (p < 0.001). We also observed significant improvements compared to Med-PaLM on every evaluation axis (p < 0.001) on newly introduced datasets of 240 long-form "adversarial" questions to probe LLM limitations. While further studies are necessary to validate the efficacy of these models in real-world settings, these results highlight rapid progress towards physician-level performance in medical question answering.
We study the problem of multi-agent control of a dynamical system with known dynamics and adversarial disturbances. Our study focuses on optimal control without centralized precomputed policies, but rather with adaptive control policies for the different agents that are only equipped with a stabilizing controller. We give a reduction from any (standard) regret minimizing control method to a distributed algorithm. The reduction guarantees that the resulting distributed algorithm has low regret relative to the optimal precomputed joint policy. Our methodology involves generalizing online convex optimization to a multi-agent setting and applying recent tools from nonstochastic control derived for a single agent. We empirically evaluate our method on a model of an overactuated aircraft. We show that the distributed method is robust to failure and to adversarial perturbations in the dynamics.
With the explosive growth of information technology, multi-view graph data have become increasingly prevalent and valuable. Most existing multi-view clustering techniques either focus on the scenario of multiple graphs or multi-view attributes. In this paper, we propose a generic framework to cluster multi-view attributed graph data. Specifically, inspired by the success of contrastive learning, we propose multi-view contrastive graph clustering (MCGC) method to learn a consensus graph since the original graph could be noisy or incomplete and is not directly applicable. Our method composes of two key steps: we first filter out the undesirable high-frequency noise while preserving the graph geometric features via graph filtering and obtain a smooth representation of nodes; we then learn a consensus graph regularized by graph contrastive loss. Results on several benchmark datasets show the superiority of our method with respect to state-of-the-art approaches. In particular, our simple approach outperforms existing deep learning-based methods.
Conversational recommender systems (CRS) aim to recommend high-quality items to users through interactive conversations. Although several efforts have been made for CRS, two major issues still remain to be solved. First, the conversation data itself lacks of sufficient contextual information for accurately understanding users' preference. Second, there is a semantic gap between natural language expression and item-level user preference. To address these issues, we incorporate both word-oriented and entity-oriented knowledge graphs (KG) to enhance the data representations in CRSs, and adopt Mutual Information Maximization to align the word-level and entity-level semantic spaces. Based on the aligned semantic representations, we further develop a KG-enhanced recommender component for making accurate recommendations, and a KG-enhanced dialog component that can generate informative keywords or entities in the response text. Extensive experiments have demonstrated the effectiveness of our approach in yielding better performance on both recommendation and conversation tasks.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191