Search

The rapid advancement of generative models in creating highly realistic images poses substantial risks for misinformation dissemination. For instance, a synthetic image, when shared on social media, can mislead extensive audiences and erode trust in digital content, resulting in severe repercussions. Despite some progress, academia has not yet created a large and diversified deepfake detection dataset for social media, nor has it devised an effective solution to address this issue. In this paper, we introduce the Social media Image Detection dataSet (SID-Set), which offers three key advantages: (1) extensive volume, featuring 300K AI-generated/tampered and authentic images with comprehensive annotations, (2) broad diversity, encompassing fully synthetic and tampered images across various classes, and (3) elevated realism, with images that are predominantly indistinguishable from genuine ones through mere visual inspection. Furthermore, leveraging the exceptional capabilities of large multimodal models, we propose a new image deepfake detection, localization, and explanation framework, named SIDA (Social media Image Detection, localization, and explanation Assistant). SIDA not only discerns the authenticity of images, but also delineates tampered regions through mask prediction and provides textual explanations of the model's judgment criteria. Compared with state-of-the-art deepfake detection models on SID-Set and other benchmarks, extensive experiments demonstrate that SIDA achieves superior performance among diversified settings. The code, model, and dataset will be released.

Generative Commonsense Reasoning (GCR) requires a model to reason about a situation using commonsense knowledge, while generating coherent sentences. Although the quality of the generated sentences is crucial, the diversity of the generation is equally important because it reflects the model's ability to use a range of commonsense knowledge facts. Large Language Models (LLMs) have shown proficiency in enhancing the generation quality across various tasks through in-context learning (ICL) using given examples without the need for any fine-tuning. However, the diversity aspect in LLM outputs has not been systematically studied before. To address this, we propose a simple method that diversifies the LLM generations, while preserving their quality. Experimental results on three benchmark GCR datasets show that our method achieves an ideal balance between the quality and diversity. Moreover, the sentences generated by our proposed method can be used as training data to improve diversity in existing commonsense generators., Comment: EMNLP 2024 Findings, Camera-ready version

ChatGPT, as a language model based on large-scale pre-training, has exerted a profound influence on the domain of machine translation. In ChatGPT, a "Prompt" refers to a segment of text or instruction employed to steer the model towards generating a specific category of response. The design of the translation prompt emerges as a key aspect that can wield influence over factors such as the style, precision and accuracy of the translation to a certain extent. However, there is a lack of a common standard and methodology on how to design and select a translation prompt. Accordingly, this paper proposes a generic taxonomy, which defines gradable translation prompts in terms of expression type, translation style, POS information and explicit statement, thus facilitating the construction of prompts endowed with distinct attributes tailored for various translation tasks. Specific experiments and cases are selected to validate and illustrate the effectiveness of the method., Comment: Publish in the journal Procesamiento del Lenguaje Natural

In the classical Kalman filter(KF), the estimated state is a linear combination of the one-step predicted state and measurement state, their confidence level change when the prediction mean square error matrix and covariance matrix of measurement noise vary. The existing student's t based Kalman filter(TKF) works similarly to the way KF works, they both work well with impulse noise, but when it comes to Gaussian noise, TKF encounters an adjustment limit of the confidence level, this can lead to inaccuracies in such situations. This brief optimizes TKF by using the Gaussian mixture model(GMM), which generates a reasonable covariance matrix from the measurement noise to replace the one used in the existing algorithm and breaks the adjustment limit of the confidence level. At the end of the brief, the performance of the covariance adaptive student's t based Kalman filter(TGKF) is verified.

In multi-target tracking (MTT), non-Gaussian measurement noise from sensors can diminish the performance of the Gaussian-assumed Gaussian mixture probability hypothesis density (GM-PHD) filter. In this paper, an approach that transforms the MTT problem under non-Gaussian conditions into an MTT problem under Gaussian conditions is developed. Specifically, measurement noise with a non-Gaussian distribution is modeled as a weighted sum of different Gaussian distributions. Subsequently, the GM-PHD filter is applied to compute the multi-target states under these distinct Gaussian distributions. Finally, an interactive multi-model framework is employed to fuse the diverse multi-target state information into a unified synthesis. The effectiveness of the proposed approach is validated through the simulation results., Comment: conference

Prior work has shown that the ordering in which concepts are shown to a commonsense generator plays an important role, affecting the quality of the generated sentence. However, it remains a challenge to determine the optimal ordering of a given set of concepts such that a natural sentence covering all the concepts could be generated from a pretrained generator. To understand the relationship between the ordering of the input concepts and the quality of the generated sentences, we conduct a systematic study considering multiple language models (LMs) and concept ordering strategies. We find that BART-large model consistently outperforms all other LMs considered in this study when fine-tuned using the ordering of concepts as they appear in CommonGen training data as measured using multiple evaluation metrics. Moreover, the larger GPT3-based large language models (LLMs) variants do not necessarily outperform much smaller LMs on this task, even when fine-tuned on task-specific training data. Interestingly, human annotators significantly reorder input concept sets when manually writing sentences covering those concepts, and this ordering provides the best sentence generations independently of the LM used for the generation, outperforming a probabilistic concept ordering baseline, Comment: 10 pages

The information transmission between nodes in a wireless sensor networks (WSNs) often causes packet loss due to denial-of-service (DoS) attack, energy limitations, and environmental factors, and the information that is successfully transmitted can also be contaminated by non-Gaussian noise. The presence of these two factors poses a challenge for distributed state estimation (DSE) over WSNs. In this paper, a generalized packet drop model is proposed to describe the packet loss phenomenon caused by DoS attacks and other factors. Moreover, a modified maximum correntropy Kalman filter is given, and it is extended to distributed form (DM-MCKF). In addition, a distributed modified maximum correntropy Kalman filter incorporating the generalized data packet drop (DM-MCKF-DPD) algorithm is provided to implement DSE with the presence of both non-Gaussian noise pollution and packet drop. A sufficient condition to ensure the convergence of the fixed-point iterative process of the DM-MCKF-DPD algorithm is presented and the computational complexity of the DM-MCKF-DPD algorithm is analyzed. Finally, the effectiveness and feasibility of the proposed algorithms are verified by simulations.

The robustness of the kernel recursive least square (KRLS) algorithm has recently been improved by combining them with more robust information-theoretic learning criteria, such as minimum error entropy (MEE) and generalized MEE (GMEE), which also improves the computational complexity of the KRLS-type algorithms to a certain extent. To reduce the computational load of the KRLS-type algorithms, the quantized GMEE (QGMEE) criterion, in this paper, is combined with the KRLS algorithm, and as a result two kinds of KRLS-type algorithms, called quantized kernel recursive MEE (QKRMEE) and quantized kernel recursive GMEE (QKRGMEE), are designed. As well, the mean error behavior, mean square error behavior, and computational complexity of the proposed algorithms are investigated. In addition, simulation and real experimental data are utilized to verify the feasibility of the proposed algorithms.

In real applications, non-Gaussian distributions are frequently caused by outliers and impulsive disturbances, and these will impair the performance of the classical cubature Kalman filter (CKF) algorithm. In this letter, a modified generalized minimum error entropy criterion with fiducial point (GMEEFP) is studied to ensure that the error comes together to around zero, and a new CKF algorithm based on the GMEEFP criterion, called GMEEFP-CKF algorithm, is developed. To demonstrate the practicality of the GMEEFP-CKF algorithm, several simulations are performed, and it is demonstrated that the proposed GMEEFP-CKF algorithm outperforms the existing CKF algorithms with impulse noise.

The distributed Kalman filter (DKF) has attracted extensive research as an information fusion method for wireless sensor systems(WSNs). And the DKF in non-Gaussian environments is still a pressing problem. In this paper, we approximate the non-Gaussian noise as a Gaussian mixture model and estimate the parameters through the expectation-maximization algorithm. A DKF, called model fusion DKF (MFDKF) is proposed against the non-Gaussain noise. Specifically, the proposed MFDKF is obtained by fusing the sub-models that are built based on the noise approximation with the help of interacting multiple model (IMM). Considering that some WSNs demand high consensus or have restricted communication, consensus MFDKF (C-MFDKF) and simplified MFDKF (S-MFDKF) are proposed based on consensus theory, respectively. The convergence of MFDKF and its derivative algorithms are analyzed. A series of simulations indicate the effectiveness of the MFDKF and its derivative algorithms.

This paper investigates how deep multi-agent reinforcement learning can enable the scalable and privacy-preserving coordination of residential energy flexibility. The coordination of distributed resources such as electric vehicles and heating will be critical to the successful integration of large shares of renewable energy in our electricity grid and, thus, to help mitigate climate change. The pre-learning of individual reinforcement learning policies can enable distributed control with no sharing of personal data required during execution. However, previous approaches for multi-agent reinforcement learning-based distributed energy resources coordination impose an ever greater training computational burden as the size of the system increases. We therefore adopt a deep multi-agent actor-critic method which uses a \emph{centralised but factored critic} to rehearse coordination ahead of execution. Results show that coordination is achieved at scale, with minimal information and communication infrastructure requirements, no interference with daily activities, and privacy protection. Significant savings are obtained for energy users, the distribution network and greenhouse gas emissions. Moreover, training times are nearly 40 times shorter than with a previous state-of-the-art reinforcement learning approach without the factored critic for 30 homes.

Distributed Kalman filter approaches based on the maximum correntropy criterion have recently demonstrated superior state estimation performance to that of conventional distributed Kalman filters for wireless sensor networks in the presence of non-Gaussian impulsive noise. However, these algorithms currently fail to take account of data packet drops. The present work addresses this issue by proposing a distributed maximum correntropy Kalman filter that accounts for data packet drops (i.e., the DMCKF-DPD algorithm). The effectiveness and feasibility of the algorithm are verified by simulations conducted in a wireless sensor network with intermittent observations due to data packet drops under a non-Gaussian noise environment. Moreover, the computational complexity of the DMCKF-DPD algorithm is demonstrated to be moderate compared with that of a conventional distributed Kalman filter, and we provide a sufficient condition to ensure the convergence of the proposed algorithm.

Outliers and impulsive disturbances often cause heavy-tailed distributions in practical applications, and these will degrade the performance of Gaussian approximation smoothing algorithms. To improve the robustness of the Rauch-Tung-Striebel (RTS) smother against complicated non-Gaussian noises, a new RTS-smoother integrated with the minimum error entropy (MEE) criterion (MEE-RTS) is proposed for linear systems, which is also extended to the state estimation of nonlinear systems by utilizing the Taylor series linearization approach. The mean error behavior, the mean square error behavior, as well as the computational complexity of the MEE-RTS smoother are analyzed. According to simulation results, the proposed smoothers perform better than several robust solutions in terms of steady-state error.

Relative overgeneralization (RO) is a pathology that can arise in cooperative multi-agent tasks when the optimal joint action's utility falls below that of a sub-optimal joint action. RO can cause the agents to get stuck into local optima or fail to solve cooperative tasks requiring significant coordination between agents within a given timestep. In this work, we empirically find that, in multi-agent reinforcement learning (MARL), both value-based and policy gradient MARL algorithms can suffer from RO and fail to learn effective coordination policies. To better overcome RO, we propose a novel approach called curriculum learning for relative overgeneralization (CURO). To solve a target task that exhibits strong RO, in CURO, we first fine-tune the reward function of the target task to generate source tasks to train the agent. Then, to effectively transfer the knowledge acquired in one task to the next, we use a transfer learning method that combines value function transfer with buffer transfer, which enables more efficient exploration in the target task. CURO is general and can be applied to both value-based and policy gradient MARL methods. We demonstrate that, when applied to QMIX, HAPPO, and HATRPO, CURO can successfully overcome severe RO, achieve improved performance, and outperform baseline methods in a variety of challenging cooperative multi-agent tasks.

The use of laboratory robotics for autonomous experiments offers an attractive route to alleviate scientists from tedious tasks while accelerating material discovery for topical issues such as climate change and pharmaceuticals. While some experimental workflows can already benefit from automation, sample preparation is still carried out manually due to the high level of motor function and dexterity required when dealing with different tools, chemicals, and glassware. A fundamental workflow in chemical fields is crystallisation, where one application is polymorph screening, i.e., obtaining a three dimensional molecular structure from a crystal. For this process, it is of utmost importance to recover as much of the sample as possible since synthesising molecules is both costly in time and money. To this aim, chemists scrape vials to retrieve sample contents prior to imaging plate transfer. Automating this process is challenging as it goes beyond robotic insertion tasks due to a fundamental requirement of having to execute fine-granular movements within a constrained environment (sample vial). Motivated by how human chemists carry out this process of scraping powder from vials, our work proposes a model-free reinforcement learning method for learning a scraping policy, leading to a fully autonomous sample scraping procedure. We first create a scenario-specific simulation environment with a Panda Franka Emika robot using a laboratory scraper that is inserted into a simulated vial, to demonstrate how a scraping policy can be learned successfully in simulation. We then train and evaluate our method on a real robotic manipulator in laboratory settings, and show that our method can autonomously scrape powder across various setups., Comment: 8 pages, 7 figures

Error entropy is a important nonlinear similarity measure, and it has received increasing attention in many practical applications. The default kernel function of error entropy criterion is Gaussian kernel function, however, which is not always the best choice. In our study, a novel concept, called generalized error entropy, utilizing the generalized Gaussian density (GGD) function as the kernel function is proposed. We further derivate the generalized minimum error entropy (GMEE) criterion, and a novel adaptive filtering called GMEE algorithm is derived by utilizing GMEE criterion. The stability, steady-state performance, and computational complexity of the proposed algorithm are investigated. Some simulation indicate that the GMEE algorithm performs well in Gaussian, sub-Gaussian, and super-Gaussian noises environment, respectively. Finally, the GMEE algorithm is applied to acoustic echo cancelation and performs well., Comment: 9 pages, 8 figures

Cooperative problems under continuous control have always been the focus of multi-agent reinforcement learning. Existing algorithms suffer from the problem of uneven learning degree with the increase of the number of agents. In this paper, a new structure for a multi-agent actor critic is proposed, and the self-attention mechanism is applied in the critic network and the value decomposition method used to solve the uneven problem. The proposed algorithm makes full use of the samples in the replay memory buffer to learn the behavior of a class of agents. First, a new update method is proposed for policy networks that promotes learning efficiency. Second, the utilization of samples is improved, at the same time reflecting the ability of perspective-taking among groups. Finally, the "deceptive signal" in training is eliminated and the learning degree among agents is more uniform than in the existing methods. Multiple experiments were conducted in two typical scenarios of a multi-agent particle environment. Experimental results show that the proposed algorithm can perform better than the state-of-the-art ones, and that it exhibits higher learning efficiency with an increasing number of agents., Comment: 30 pages

Policy gradient methods are an attractive approach to multi-agent reinforcement learning problems due to their convergence properties and robustness in partially observable scenarios. However, there is a significant performance gap between state-of-the-art policy gradient and value-based methods on the popular StarCraft Multi-Agent Challenge (SMAC) benchmark. In this paper, we introduce semi-on-policy (SOP) training as an effective and computationally efficient way to address the sample inefficiency of on-policy policy gradient methods. We enhance two state-of-the-art policy gradient algorithms with SOP training, demonstrating significant performance improvements. Furthermore, we show that our methods perform as well or better than state-of-the-art value-based methods on a variety of SMAC tasks., Comment: AAMAS Adaptive and Learning Agents Workshop. 20th International Conference on Autonomous Agents and Multiagent Systems

Tackling overestimation in $Q$-learning is an important problem that has been extensively studied in single-agent reinforcement learning, but has received comparatively little attention in the multi-agent setting. In this work, we empirically demonstrate that QMIX, a popular $Q$-learning algorithm for cooperative multi-agent reinforcement learning (MARL), suffers from a more severe overestimation in practice than previously acknowledged, and is not mitigated by existing approaches. We rectify this with a novel regularization-based update scheme that penalizes large joint action-values that deviate from a baseline and demonstrate its effectiveness in stabilizing learning. Furthermore, we propose to employ a softmax operator, which we efficiently approximate in a novel way in the multi-agent setting, to further reduce the potential overestimation bias. Our approach, Regularized Softmax (RES) Deep Multi-Agent $Q$-Learning, is general and can be applied to any $Q$-learning based MARL algorithm. We demonstrate that, when applied to QMIX, RES avoids severe overestimation and significantly improves performance, yielding state-of-the-art results in a variety of cooperative multi-agent tasks, including the challenging StarCraft II micromanagement benchmarks.

VDN and QMIX are two popular value-based algorithms for cooperative MARL that learn a centralized action value function as a monotonic mixing of per-agent utilities. While this enables easy decentralization of the learned policy, the restricted joint action value function can prevent them from solving tasks that require significant coordination between agents at a given timestep. We show that this problem can be overcome by improving the joint exploration of all agents during training. Specifically, we propose a novel MARL approach called Universal Value Exploration (UneVEn) that learns a set of related tasks simultaneously with a linear decomposition of universal successor features. With the policies of already solved related tasks, the joint exploration process of all agents can be improved to help them achieve better coordination. Empirical results on a set of exploration games, challenging cooperative predator-prey tasks requiring significant coordination among agents, and StarCraft II micromanagement benchmarks show that UneVEn can solve tasks where other state-of-the-art MARL methods fail., Comment: Published at ICML 2021

Role-based learning holds the promise of achieving scalable multi-agent learning by decomposing complex tasks using roles. However, it is largely unclear how to efficiently discover such a set of roles. To solve this problem, we propose to first decompose joint action spaces into restricted role action spaces by clustering actions according to their effects on the environment and other agents. Learning a role selector based on action effects makes role discovery much easier because it forms a bi-level learning hierarchy -- the role selector searches in a smaller role space and at a lower temporal resolution, while role policies learn in significantly reduced primitive action-observation spaces. We further integrate information about action effects into the role policies to boost learning efficiency and policy generalization. By virtue of these advances, our method (1) outperforms the current state-of-the-art MARL algorithms on 10 of the 14 scenarios that comprise the challenging StarCraft II micromanagement benchmark and (2) achieves rapid transfer to new environments with three times the number of agents. Demonstrative videos are available at https://sites.google.com/view/rode-marl .

QMIX is a popular $Q$-learning algorithm for cooperative MARL in the centralised training and decentralised execution paradigm. In order to enable easy decentralisation, QMIX restricts the joint action $Q$-values it can represent to be a monotonic mixing of each agent's utilities. However, this restriction prevents it from representing value functions in which an agent's ordering over its actions can depend on other agents' actions. To analyse this representational limitation, we first formalise the objective QMIX optimises, which allows us to view QMIX as an operator that first computes the $Q$-learning targets and then projects them into the space representable by QMIX. This projection returns a representable $Q$-value that minimises the unweighted squared error across all joint actions. We show in particular that this projection can fail to recover the optimal policy even with access to $Q^*$, which primarily stems from the equal weighting placed on each joint action. We rectify this by introducing a weighting into the projection, in order to place more importance on the better joint actions. We propose two weighting schemes and prove that they recover the correct maximal action for any joint action $Q$-values, and therefore for $Q^*$ as well. Based on our analysis and results in the tabular setting, we introduce two scalable versions of our algorithm, Centrally-Weighted (CW) QMIX and Optimistically-Weighted (OW) QMIX and demonstrate improved performance on both predator-prey and challenging multi-agent StarCraft benchmark tasks.

Multi-agent settings in the real world often involve tasks with varying types and quantities of agents and non-agent entities; however, common patterns of behavior often emerge among these agents/entities. Our method aims to leverage these commonalities by asking the question: ``What is the expected utility of each agent when only considering a randomly selected sub-group of its observed entities?'' By posing this counterfactual question, we can recognize state-action trajectories within sub-groups of entities that we may have encountered in another task and use what we learned in that task to inform our prediction in the current one. We then reconstruct a prediction of the full returns as a combination of factors considering these disjoint groups of entities and train this ``randomly factorized" value function as an auxiliary objective for value-based multi-agent reinforcement learning. By doing so, our model can recognize and leverage similarities across tasks to improve learning efficiency in a multi-task setting. Our approach, Randomized Entity-wise Factorization for Imagined Learning (REFIL), outperforms all strong baselines by a significant margin in challenging multi-task StarCraft micromanagement settings., Comment: ICML 2021 Camera Ready

We propose FACtored Multi-Agent Centralised policy gradients (FACMAC), a new method for cooperative multi-agent reinforcement learning in both discrete and continuous action spaces. Like MADDPG, a popular multi-agent actor-critic method, our approach uses deep deterministic policy gradients to learn policies. However, FACMAC learns a centralised but factored critic, which combines per-agent utilities into the joint action-value function via a non-linear monotonic function, as in QMIX, a popular multi-agent Q-learning algorithm. However, unlike QMIX, there are no inherent constraints on factoring the critic. We thus also employ a nonmonotonic factorisation and empirically demonstrate that its increased representational capacity allows it to solve some tasks that cannot be solved with monolithic, or monotonically factored critics. In addition, FACMAC uses a centralised policy gradient estimator that optimises over the entire joint action space, rather than optimising over each agent's action space separately as in MADDPG. This allows for more coordinated policy changes and fully reaps the benefits of a centralised critic. We evaluate FACMAC on variants of the multi-agent particle environments, a novel multi-agent MuJoCo benchmark, and a challenging set of StarCraft II micromanagement tasks. Empirical results demonstrate FACMAC's superior performance over MADDPG and other baselines on all three domains.