In this work, we study to release the potential of massive heterogeneous weak computing power to collaboratively train large-scale models on dispersed datasets. In order to improve both efficiency and accuracy in resource-adaptive collaborative learning, we take the first step to consider the \textit{unstructured pruning}, \textit{varying submodel architectures}, \textit{knowledge loss}, and \textit{straggler} challenges simultaneously. We propose a novel semi-asynchronous collaborative training framework, namely ${Co\text{-}S}^2{P}$, with data distribution-aware structured pruning and cross-block knowledge transfer mechanism to address the above concerns. Furthermore, we provide theoretical proof that ${Co\text{-}S}^2{P}$ can achieve asymptotic optimal convergence rate of $O(1/\sqrt{N^*EQ})$. Finally, we conduct extensive experiments on a real-world hardware testbed, in which 16 heterogeneous Jetson devices can be united to train large-scale models with parameters up to 0.11 billion. The experimental results demonstrate that $Co\text{-}S^2P$ improves accuracy by up to 8.8\% and resource utilization by up to 1.2$\times$ compared to state-of-the-art methods, while reducing memory consumption by approximately 22\% and training time by about 24\% on all resource-limited devices.

The task that requires an agent to navigate to a given object through only visual observation is called visual object navigation (VON). The main bottlenecks of VON are strategies exploration and prior knowledge exploitation. Traditional strategies exploration ignores the differences of searching and navigating stages, using the same reward in two stages, which reduces navigation performance and training efficiency. Our study enables the agent to explore larger area in searching stage and seek the optimal path in navigating stage, improving the success rate of navigation. Traditional prior knowledge exploitation focused on learning and utilizing object association, which ignored the depth and obstacle information in the environment. This paper uses the RGB and depth information of the training scene to pretrain the feature extractor, which improves navigation efficiency. The obstacle information is memorized by the agent during the navigation, reducing the probability of collision and deadlock. Depth, obstacle and other prior knowledge are concatenated and input into the policy network, and navigation actions are output under the training of two-stage rewards. We evaluated our method on AI2-Thor and RoboTHOR and demonstrated that it significantly outperforms state-of-the-art (SOTA) methods on success rate and navigation efficiency.

Zero-shot object navigation (ZSON) addresses situation where an agent navigates to an unseen object that does not present in the training set. Previous works mainly train agent using seen objects with known labels, and ignore the seen objects without labels. In this paper, we introduce seen objects without labels, herein termed as ``unknown objects'', into training procedure to enrich the agent's knowledge base with distinguishable but previously overlooked information. Furthermore, we propose the label-wise meta-correlation module (LWMCM) to harness relationships among objects with and without labels, and obtain enhanced objects information. Specially, we propose target feature generator (TFG) to generate the features representation of the unlabeled target objects. Subsequently, the unlabeled object identifier (UOI) module assesses whether the unlabeled target object appears in the current observation frame captured by the camera and produces an adapted target features representation specific to the observed context. In meta contrastive feature modifier (MCFM), the target features is modified via approaching the features of objects within the observation frame while distancing itself from features of unobserved objects. Finally, the meta object-graph learner (MOGL) module is utilized to calculate the relationships among objects based on the features. Experiments conducted on AI2THOR and RoboTHOR platforms demonstrate the effectiveness of our proposed method.

The safety of decentralized reinforcement learning (RL) is a challenging problem since malicious agents can share their poisoned policies with benign agents. The paper investigates a cooperative backdoor attack in a decentralized reinforcement learning scenario. Differing from the existing methods that hide a whole backdoor attack behind their shared policies, our method decomposes the backdoor behavior into multiple components according to the state space of RL. Each malicious agent hides one component in its policy and shares its policy with the benign agents. When a benign agent learns all the poisoned policies, the backdoor attack is assembled in its policy. The theoretical proof is given to show that our cooperative method can successfully inject the backdoor into the RL policies of benign agents. Compared with the existing backdoor attacks, our cooperative method is more covert since the policy from each attacker only contains a component of the backdoor attack and is harder to detect. Extensive simulations are conducted based on Atari environments to demonstrate the efficiency and covertness of our method. To the best of our knowledge, this is the first paper presenting a provable cooperative backdoor attack in decentralized reinforcement learning.

We study federated unlearning, a novel problem to eliminate the impact of specific clients or data points on the global model learned via federated learning (FL). This problem is driven by the right to be forgotten and the privacy challenges in FL. We introduce a new framework for exact federated unlearning that meets two essential criteria: \textit{communication efficiency} and \textit{exact unlearning provability}. To our knowledge, this is the first work to tackle both aspects coherently. We start by giving a rigorous definition of \textit{exact} federated unlearning, which guarantees that the unlearned model is statistically indistinguishable from the one trained without the deleted data. We then pinpoint the key property that enables fast exact federated unlearning: total variation (TV) stability, which measures the sensitivity of the model parameters to slight changes in the dataset. Leveraging this insight, we develop a TV-stable FL algorithm called \texttt{FATS}, which modifies the classical \texttt{\underline{F}ed\underline{A}vg} algorithm for \underline{T}V \underline{S}tability and employs local SGD with periodic averaging to lower the communication round. We also design efficient unlearning algorithms for \texttt{FATS} under two settings: client-level and sample-level unlearning. We provide theoretical guarantees for our learning and unlearning algorithms, proving that they achieve exact federated unlearning with reasonable convergence rates for both the original and unlearned models. We empirically validate our framework on 6 benchmark datasets, and show its superiority over state-of-the-art methods in terms of accuracy, communication cost, computation cost, and unlearning efficacy.

Federated learning is a powerful technique that enables collaborative learning among different clients. Prototype-based federated learning is a specific approach that improves the performance of local models under non-IID (non-Independently and Identically Distributed) settings by integrating class prototypes. However, prototype-based federated learning faces several challenges, such as prototype redundancy and prototype failure, which limit its accuracy. It is also susceptible to poisoning attacks and server malfunctions, which can degrade the prototype quality. To address these issues, we propose FedRFQ, a prototype-based federated learning approach that aims to reduce redundancy, minimize failures, and improve \underline{q}uality. FedRFQ leverages a SoftPool mechanism, which effectively mitigates prototype redundancy and prototype failure on non-IID data. Furthermore, we introduce the BFT-detect, a BFT (Byzantine Fault Tolerance) detectable aggregation algorithm, to ensure the security of FedRFQ against poisoning attacks and server malfunctions. Finally, we conduct experiments on three different datasets, namely MNIST, FEMNIST, and CIFAR-10, and the results demonstrate that FedRFQ outperforms existing baselines in terms of accuracy when handling non-IID data.

Value function factorization has achieved great success in multi-agent reinforcement learning by optimizing joint action-value functions through the maximization of factorized per-agent utilities. To ensure Individual-Global-Maximum property, existing works often focus on value factorization using monotonic functions, which are known to result in restricted representation expressiveness. In this paper, we analyze the limitations of monotonic factorization and present ConcaveQ, a novel non-monotonic value function factorization approach that goes beyond monotonic mixing functions and employs neural network representations of concave mixing functions. Leveraging the concave property in factorization, an iterative action selection scheme is developed to obtain optimal joint actions during training. It is used to update agents' local policy networks, enabling fully decentralized execution. The effectiveness of the proposed ConcaveQ is validated across scenarios involving multi-agent predator-prey environment and StarCraft II micromanagement tasks. Empirical results exhibit significant improvement of ConcaveQ over state-of-the-art multi-agent reinforcement learning approaches.

Distributed stochastic optimization methods based on Newton's method offer significant advantages over first-order methods by leveraging curvature information for improved performance. However, the practical applicability of Newton's method is hindered in large-scale and heterogeneous learning environments due to challenges such as high computation and communication costs associated with the Hessian matrix, sub-model diversity, staleness in training, and data heterogeneity. To address these challenges, this paper introduces a novel and efficient algorithm called RANL, which overcomes the limitations of Newton's method by employing a simple Hessian initialization and adaptive assignments of training regions. The algorithm demonstrates impressive convergence properties, which are rigorously analyzed under standard assumptions in stochastic optimization. The theoretical analysis establishes that RANL achieves a linear convergence rate while effectively adapting to available resources and maintaining high efficiency. Unlike traditional first-order methods, RANL exhibits remarkable independence from the condition number of the problem and eliminates the need for complex parameter tuning. These advantages make RANL a promising approach for distributed stochastic optimization in practical scenarios.

We consider a K-armed bandit problem in general graphs where agents are arbitrarily connected and each of them has limited memorizing capabilities and communication bandwidth. The goal is to let each of the agents eventually learn the best arm. It is assumed in these studies that the communication graph should be complete or well-structured, whereas such an assumption is not always valid in practice. Furthermore, limited memorization and communication bandwidth also restrict the collaborations of the agents, since the agents memorize and communicate very few experiences. Additionally, an agent may be corrupted to share falsified experiences to its peers, while the resource limit in terms of memorization and communication may considerably restrict the reliability of the learning process. To address the above issues, we propose a three-staged collaborative learning algorithm. In each step, the agents share their latest experiences with each other through light-weight random walks in a general communication graph, and then make decisions on which arms to pull according to the recommendations received from their peers. The agents finally update their adoptions (i.e., preferences to the arms) based on the reward obtained by pulling the arms. Our theoretical analysis shows that, when there are a sufficient number of agents participating in the collaborative learning process, all the agents eventually learn the best arm with high probability, even with limited memorizing capabilities and light-weight communications. We also reveal in our theoretical analysis the upper bound on the number of corrupted agents our algorithm can tolerate. The efficacy of our proposed three-staged collaborative learning algorithm is finally verified by extensive experiments on both synthetic and real datasets.

Both Byzantine resilience and communication efficiency have attracted tremendous attention recently for their significance in edge federated learning. However, most existing algorithms may fail when dealing with real-world irregular data that behaves in a heavy-tailed manner. To address this issue, we study the stochastic convex and non-convex optimization problem for federated learning at edge and show how to handle heavy-tailed data while retaining the Byzantine resilience, communication efficiency and the optimal statistical error rates simultaneously. Specifically, we first present a Byzantine-resilient distributed gradient descent algorithm that can handle the heavy-tailed data and meanwhile converge under the standard assumptions. To reduce the communication overhead, we further propose another algorithm that incorporates gradient compression techniques to save communication costs during the learning process. Theoretical analysis shows that our algorithms achieve order-optimal statistical error rate in presence of Byzantine devices. Finally, we conduct extensive experiments on both synthetic and real-world datasets to verify the efficacy of our algorithms.