Optimization in the latent space of variational autoencoders is a promising approach to generate high-dimensional discrete objects that maximize an expensive black-box property (e.g., drug-likeness in molecular generation, function approximation with arithmetic expressions). However, existing methods lack robustness as they may decide to explore areas of the latent space for which no data was available during training and where the decoder can be unreliable, leading to the generation of unrealistic or invalid objects. We propose to leverage the epistemic uncertainty of the decoder to guide the optimization process. This is not trivial though, as a naive estimation of uncertainty in the high-dimensional and structured settings we consider would result in high estimator variance. To solve this problem, we introduce an importance sampling-based estimator that provides more robust estimates of epistemic uncertainty. Our uncertainty-guided optimization approach does not require modifications of the model architecture nor the training process. It produces samples with a better trade-off between black-box objective and validity of the generated samples, sometimes improving both simultaneously. We illustrate these advantages across several experimental settings in digit generation, arithmetic expression approximation and molecule generation for drug design.

Neural networks are known to be susceptible to adversarial samples: small variations of natural examples crafted to deliberately mislead the models. While they can be easily generated using gradient-based techniques in digital and physical scenarios, they often differ greatly from the actual data distribution of natural images, resulting in a trade-off between strength and stealthiness. In this paper, we propose a novel framework dubbed Diffusion-Based Projected Gradient Descent (Diff-PGD) for generating realistic adversarial samples. By exploiting a gradient guided by a diffusion model, Diff-PGD ensures that adversarial samples remain close to the original data distribution while maintaining their effectiveness. Moreover, our framework can be easily customized for specific tasks such as digital attacks, physical-world attacks, and style-based attacks. Compared with existing methods for generating natural-style adversarial samples, our framework enables the separation of optimizing adversarial loss from other surrogate losses (e.g., content/smoothness/style loss), making it more stable and controllable. Finally, we demonstrate that the samples generated using Diff-PGD have better transferability and anti-purification power than traditional gradient-based methods.

The fuzzy or soft k-means objective is a popular generalization of the well-known kmeans problem, extending the clustering capability of the k-means to datasets that are uncertain, vague and otherwise hard to cluster. In this paper, we propose a semisupervised active clustering framework, where the learner is allowed to interact with an oracle (domain expert), asking for the similarity between a certain set of chosen items. We study the query and computational complexities of clustering in this framework. We prove that having a few of such similarity queries enables one to get a polynomial-time approximation algorithm to an otherwise conjecturally NP-hard problem. In particular, we provide algorithms for fuzzy clustering in this setting that ask O(poly(k)logn)similarity queries and run with polynomialtime-complexity, where nis the number of items. The fuzzy k-means objective is nonconvex, with k-means as a special case, and is equivalent to some other generic nonconvex problem such as non-negative matrix factorization. The ubiquitous Lloyd-type algorithms (or alternating-minimization algorithms) can get stuck at a local minima. Our results show that by making few similarity queries, the problem becomes easier to solve. Finally, we test our algorithms over real-world datasets, showing their effectiveness in real-world applications.

Offline reinforcement learning (RL) defines the task of learning from a static logged dataset without continually interacting with the environment. The distribution shift between the learned policy and the behavior policy makes it necessary for the value function to stay conservative such that out-of-distribution (OOD) actions will not be severely overestimated. However, existing approaches, penalizing the unseen actions or regularizing with the behavior policy, are too pessimistic, which suppresses the generalization of the value function and hinders the performance improvement. This paper explores mild but enough conservatism for offline learning while not harming generalization. We propose Mildly Conservative Q-learning (MCQ), where OOD actions are actively trained by assigning them proper pseudo Qvalues. We theoretically show that MCQ induces a policy that behaves at least as well as the behavior policy and no erroneous overestimation will occur for OOD actions. Experimental results on the D4RL benchmarks demonstrate that MCQ achieves remarkable performance compared with prior work. Furthermore, MCQ shows superior generalization ability when transferring from offline to online, and significantly outperforms baselines. Our code is publicly available at https://github.com/dmksjfl/MCQ.

Cooperative multi-agent reinforcement learning (MARL) has made prominent progress in recent years. For training efficiency and scalability, most of the MARL algorithms make all agents share the same policy or value network. However, in many complex multi-agent tasks, different agents are expected to possess specific abilities to handle different subtasks. In those scenarios, sharing parameters indiscriminately may lead to similar behavior across all agents, which will limit the exploration efficiency and degrade the final performance. To balance the training complexity and the diversity of agent behavior, we propose a novel framework to learn dynamic subtask assignment (LDSA) in cooperative MARL. Specifically, we first introduce a subtask encoder to construct a vector representation for each subtask according to its identity.

Modeling distributions of covariates, or density estimation, is a core challenge in unsupervised learning. However, the majority of work only considers the joint distribution, which has limited utility in practical situations. A more general and useful problem is arbitrary conditional density estimation, which aims to model any possible conditional distribution over a set of covariates, reflecting the more realistic setting of inference based on prior knowledge. We propose a novel method, Arbitrary Conditioning with Energy (ACE), that can simultaneously estimate the distribution p(xu | xo) for all possible subsets of unobserved features xu and observed features xo. ACE is designed to avoid unnecessary bias and complexity -- we specify densities with a highly expressive energy function and reduce the problem to only learning one-dimensional conditionals (from which more complex distributions can be recovered during inference). This results in an approach that is both simpler and higher-performing than prior methods. We show that ACE achieves state-of-the-art for arbitrary conditional likelihood estimation and data imputation on standard benchmarks.

Sequential recommender systems (SRSs) are typically trained to predict the next item as the target given its preceding (and succeeding) items as the input. Such a paradigm assumes that every input-target pair is reliable for training. However, users can be induced to click on items that are inconsistent with their true preferences, resulting in unreliable instances, i.e., mismatched input-target pairs. Current studies on mitigating this issue suffer from two limitations: (i) they discriminate instance reliability according to models trained with unreliable data, yet without theoretical guarantees that such a seemingly contradictory solution can be effective; and (ii) most methods can only tackle either unreliable input or targets but fail to handle both simultaneously. To fill the gap, we theoretically unveil the relationship between SRS predictions and instance reliability, whereby two error-bounded strategies are proposed to rectify unreliable targets and input, respectively. On this basis, we devise a model-agnostic Bidirectional Data Rectification (BirDRec) framework, which can be flexibly implemented with most existing SRSs for robust training against unreliable data. Additionally, a rectification sampling strategy is devised and a self-ensemble mechanism is adopted to reduce the (time and space) complexity of BirDRec. Extensive experiments on four real-world datasets verify the generality, effectiveness, and efficiency of our proposed BirDRec.