Diffusion language models (DLMs) have recently emerged as a promising alternative to autoregressive (AR) approaches, enabling parallel token generation beyond a rigid left-to-right order. Despite growing empirical success, the theoretical understanding of how unmasking schedules -- which specify the order and size of unmasked tokens during sampling -- affect generation quality remains limited. In this work, we introduce a distribution-agnostic unmasking schedule for DLMs that adapts to the (unknown) dependence structure of the target data distribution, without requiring any prior knowledge or hyperparameter tuning. In contrast to prior deterministic procedures that fix unmasking sizes, our method randomizes the number of tokens revealed at each iteration. We show that, for two specific parameter choices, the sampling convergence guarantees -- measured by Kullback-Leibler (KL) divergence -- scale as $\widetilde O(\mathsf{TC}/K)$ and $\widetilde O(\mathsf{DTC}/K)$ respectively. Here, $K$ is the number of iterations, and $\mathsf{TC}$ and $\mathsf{DTC}$ are the total correlation and dual total correlation of the target distribution, capturing the intrinsic dependence structure underlying the data. Importantly, our guarantees hold in the practically relevant parallel-sampling regime $K

Neural Information Processing Systems http://nips.cc/

Neural Information Processing Systems http://nips.cc/

Despite the success in learning semantically meaningful, unsupervised disentangled representations, variational autoencoders (VAEs) and their variants face a fundamental theoretical challenge: substantial evidence indicates that unsupervised disentanglement is unattainable without implicit inductive bias, yet such bias remains elusive. In this work, we focus on exploring the implicit inductive bias that drive disentanglement in VAEs with factorization priors. By analyzing the total correlation in \b{eta}-TCVAE, we uncover a crucial implicit inductive bias called disentangling granularity, which leads to the discovery of an interesting "V"-shaped optimal Evidence Lower Bound (ELBO) trajectory within the parameter space. This finding is validated through over 100K experiments using factorized VAEs and our newly proposed model, \b{eta}-STCVAE. Notably, experimental results reveal that conventional factorized VAEs, constrained by fixed disentangling granularity, inherently tend to disentangle low-complexity feature. Whereas, appropriately tuning disentangling granularity, as enabled by \b{eta}-STCVAE, broadens the range of disentangled representations, allowing for the disentanglement of high-complexity features. Our findings unveil that disentangling granularity as an implicit inductive bias in factorized VAEs influence both disentanglement performance and the inference of the ELBO, offering fresh insights into the interpretability and inherent biases of VAEs.

Simplicity is a powerful inductive bias. In reinforcement learning, regularization is used for simpler policies, data augmentation for simpler representations, and sparse reward functions for simpler objectives, all that, with the underlying motivation to increase generalizability and robustness by focusing on the essentials. Supplementary to these techniques, we investigate how to promote simple behavior throughout the episode. To that end, we introduce a modification of the reinforcement learning problem that additionally maximizes the total correlation within the induced trajectories. We propose a practical algorithm that optimizes all models, including policy and state representation, based on a lower-bound approximation. In simulated robot environments, our method naturally generates policies that induce periodic and compressible trajectories, and that exhibit superior robustness to noise and changes in dynamics compared to baseline methods, while also improving performance in the original tasks.

We study stochastic multi-agent systems in which agents must cooperate to maximize the probability of achieving a common reach-avoid objective. In many applications, during the execution of the system, the communication between the agents can be constrained by restrictions on the bandwidth currently available for exchanging local-state information between the agents. In this paper, we propose a method for computing joint action and communication policies for the group of agents that aim to satisfy the communication restrictions as much as possible while achieving the optimal reach-avoid probability when communication is unconstrained. Our method synthesizes a pair of action and communication policies robust to restrictions on the number of agents allowed to communicate. To this end, we introduce a novel cost function that measures the amount of information exchanged beyond what the communication policy allows. We evaluate our approach experimentally on a range of benchmarks and demonstrate that it is capable of computing pairs of action and communication policies that satisfy the communication restrictions, if such exist.