Modeling the spatio-temporal evolution of complex physical sys- tems remains a fundamental challenge in both deep learning and sci- entific computing. While recent methods such as Transformers and Neural Operators have shown promise in learning PDE solutions, their reliance on auto-regressive forecasting often increases compu- tational overhead and accumulates prediction errors over time. In this paper, we propose Learnable Orthogonal Decomposition (LOD), a non-regressive framework that integrates ideas from classical Proper Orthogonal Decomposition (POD) with modern deep learn- ing. LOD first performs parameter-wise POD: at each time step, we apply POD to an ensemble of PDE solutions generated under differ- ent physical parameters, yielding a time-indexed set of orthonormal spatial bases. These bases initialize a learnable dictionary and are refined during end-to-end training. Given only a short prefix of ini- tial conditions for a target parameter setting, a neural encoder pre- dicts-in a single shot-the entire trajectory of parameter-wise POD coefficients. The final solution field is reconstructed by combining the predicted coefficients with the learned bases, avoiding error ac- cumulation inherent to auto-regressive strategies. Comprehensive experiments on various PDE benchmark datasets demonstrate that LOD achieves state-of-the-art accuracy while significantly reducing computational costs.

Variational inference often struggles with the posterior geometry exhibited by complex hierarchical Bayesian models. Recent advances in flow‐based variational families and Variationally Inferred Parameters (VIP) each address aspects of this challenge, but their formal relationship is unexplored. Here, we prove that the combination of VIP and a full-rank Gaussian can be represented exactly as a forward autoregressive flow augmented with a translation term and input from the model’s prior. Guided by this theoretical insight, we introduce the Model‐Informed Flow (MIF) architecture, which adds the necessary translation mechanism, prior information, and hierarchical ordering. Empirically, MIF delivers tighter posterior approximations and matches or exceeds state‐of‐the‐art performance across a suite of hierarchical and non‐hierarchical benchmarks.

Optimization objectives in the form of a sum of intractable expectations are rising in importance (e.g.,, diffusion models, variational autoencoders, and many more), a setting also known as "finite sum with infinite data." For these problems, a popular strategy is to employ SGD with doubly stochastic gradients (doubly SGD): the expectations are estimated using the gradient estimator of each component, while the sum is estimated by subsampling over these estimators. Despite its popularity, little is known about the convergence properties of doubly SGD, except under strong assumptions such as bounded variance. In this work, we establish the convergence of doubly SGD with independent minibatching and random reshuffling under general conditions, which encompasses dependent component gradient estimators. In particular, for dependent estimators, our analysis allows fined-grained analysis of the effect correlations. As a result, under a per-iteration computational budget of $b \times m$, where $b$ is the minibatch size and $m$ is the number of Monte Carlo samples, our analysis suggests where one should invest most of the budget in general. Furthermore, we prove that random reshuffling (RR) improves the complexity dependence on the subsampling noise.

GFlowNets are probabilistic models that sequentially generate compositional structures through a stochastic policy. Among GFlowNets, temperature-conditional GFlowNets can introduce temperature-based controllability for exploration and exploitation. We propose Logit-scaling GFlowNets (Logit-GFN), a novel architectural design that greatly accelerates the training of temperature-conditional GFlowNets. It is based on the idea that previously proposed approaches introduced numerical challenges in the deep network training, since different temperatures may give rise to very different gradient profiles as well as magnitudes of the policy’s logits. We find that the challenge is greatly reduced if a learned function of the temperature is used to scale the policy’s logits directly. Also, using Logit-GFN, GFlowNets can be improved by having better generalization capabilities in offline learning and mode discovery capabilities in online learning, which is empirically verified in various biological and chemical tasks. Our code is available at https://github.com/dbsxodud-11/logit-gfn

Variational families with full-rank covariance approximations are known not to work well in black-box variational inference (BBVI), both empirically and theoretically. In fact, recent computational complexity results for BBVI have established that full-rank variational families scale poorly with the dimensionality of the problem compared to *e.g.* mean-field families. This is particularly critical to hierarchical Bayesian models with local variables; their dimensionality increases with the size of the datasets. Consequently, one gets an iteration complexity with an explicit $\mathcal{O}(N^2)$ dependence on the dataset size $N$. In this paper, we explore a theoretical middle ground *between* mean-field variational families and full-rank families: *structured* variational families. We rigorously prove that certain scale matrix structures can achieve a better iteration complexity of $\mathcal{O}\left(N\right)$, implying better scaling with respect to $N$. We empirically verify our theoretical results on large-scale hierarchical models.

Due to the lack of state dimension optimization methods, deep state space models (SSMs) have sacrificed model capacity, training search space, or stability to alleviate computational costs caused by high state dimensions. In this work, we provide a structured pruning method for SSMs, Layer-Adaptive STate pruning (LAST), which reduces the state dimension of each layer in minimizing model-level output energy loss by extending modal truncation for a single system. LAST scores are evaluated using the $\mathcal{H}_{\infty}$ norms of subsystems and layer-wise energy normalization. The scores serve as global pruning criteria, enabling cross-layer comparison of states and layer-adaptive pruning. Across various sequence benchmarks, LAST optimizes previous SSMs, revealing the redundancy and compressibility of their state spaces. Notably, we demonstrate that, on average, pruning 33\% of states still maintains performance with 0.52\% accuracy loss in multi-input multi-output SSMs without retraining. Code is available at https://github.com/msgwak/LAST.