Approximate Thompson sampling with Langevin Monte Carlo broadens its reach from Gaussian posterior sampling to encompass more general smooth posteriors. However, it still encounters scalability issues in high-dimensional problems when demanding high accuracy. To address this, we propose an approximate Thompson sampling strategy, utilizing underdamped Langevin Monte Carlo, where the latter is the go-to workhorse for simulations of high-dimensional posteriors. Based on the standard smoothness and log-concavity conditions, we study the accelerated posterior concentration and sampling using a specific potential function. This design improves the sample complexity for realizing logarithmic regrets from $\mathcal{\tilde O}(d)$ to $\mathcal{\tilde O}(\sqrt{d})$. The scalability and robustness of our algorithm are also empirically validated through synthetic experiments in high-dimensional bandit problems.

Nonparametric models have extensive applications across diverse fields, including biology (Mac-Farland et al. (2016)), economics (Ullah and Pagan (1999); Li and Racine (2023)), engineering (Lanzante (1996)), and machine learning (Hofmann et al. (2008); Schmidt-hieber (2020)). The most representative nonparametric approaches are kernel methods, known for their numerical robustness and statistical stability in lower-dimensional settings. However, kernel methods often suffer from the curse of dimensionality in higher-dimensional spaces. Recently, a number of significant studies have tackled various modern challenges in nonparametric models. For example, Ravikumar et al. (2009), Raskutti et al. (2012), and Yuan and Zhou (2016) have studied additive models for high-dimensional nonparametric regression; Zhang et al. (2015) and Yang et al. (2017) analyzed randomized algorithms for kernel regression estimation; and Liu et al. (2007) explored nonparametric density estimation in higher dimensions. Despite these contributions, the curse of dimensionality in nonparametric problems, particularly in aspects of statistical accuracy and computational efficiency, remains an open area for further exploration. In this paper, we aim to develop a new framework specifically designed for nonparametric estimation problems. Within this framework, we conceptualize functions as matrices or tensors and explore new methods for handling the bias-variance trade-off, aiming to reduce the curse of dimensionality in higher dimensions. Matrix approximation algorithms, such as singular value decomposition and QR decomposition, play a crucial role in computational mathematics and statistics.

Even with several advancements in multilingual modeling, it is challenging to recognize multiple languages using a single neural model, without knowing the input language and most multilingual models assume the availability of the input language. In this work, we propose a novel bilingual end-to-end (E2E) modeling approach, where a single neural model can recognize both languages and also support switching between the languages, without any language input from the user. The proposed model has shared encoder and prediction networks, with language-specific joint networks that are combined via a self-attention mechanism. As the language-specific posteriors are combined, it produces a single posterior probability over all the output symbols, enabling a single beam search decoding and also allowing dynamic switching between the languages. The proposed approach outperforms the conventional bilingual baseline with 13.3%, 8.23% and 1.3% word error rate relative reduction on Hindi, English and code-mixed test sets, respectively.

The study proposes the reformulation of offline reinforcement learning as a regression problem that can be solved with decision trees. Aiming to predict actions based on input states, return-to-go (RTG), and timestep information, we observe that with gradient-boosted trees, the agent training and inference are very fast, the former taking less than a minute. Despite the simplification inherent in this reformulated problem, our agent demonstrates performance that is at least on par with established methods. This assertion is validated by testing it across standard datasets associated with D4RL Gym-MuJoCo tasks. We further discuss the agent's ability to generalize by testing it on two extreme cases, how it learns to model the return distributions effectively even with highly skewed expert datasets, and how it exhibits robust performance in scenarios with sparse/delayed rewards.

Why do neural networks trained with large learning rates for a longer time often lead to better generalization? In this paper, we delve into this question by examining the relation between training and testing loss in neural networks. Through visualization of these losses, we note that the training trajectory with a large learning rate navigates through the minima manifold of the training loss, finally nearing the neighborhood of the testing loss minimum. Motivated by these findings, we introduce a nonlinear model whose loss landscapes mirror those observed for real neural networks. Upon investigating the training process using SGD on our model, we demonstrate that an extended phase with a large learning rate steers our model towards the minimum norm solution of the training loss, which may achieve near-optimal generalization, thereby affirming the empirically observed benefits of late learning rate decay.

Utilizing unsupervised representation learning for quantum architecture search (QAS) represents a cutting-edge approach poised to realize potential quantum advantage on Noisy Intermediate-Scale Quantum (NISQ) devices. QAS is a scheme to design quantum circuits for variational quantum algorithms (VQAs). Most QAS algorithms combine their search space and search algorithms together and thus generally require evaluating a large number of quantum circuits during the search process, which results in formidable computational demands and limits their applications to large-scale quantum circuits. Predictor-based QAS algorithms can alleviate this problem by directly estimating the performance of circuits according to their structures. However, a high-performance predictor generally requires very time-consuming labeling to obtain a large number of labeled quantum circuits because the gate parameters of quantum circuits need to be optimized until convergence to their ground-truth performances. Recently, a classical neural architecture search algorithm Arch2vec inspires us by showing that architecture search can benefit from decoupling unsupervised representation learning from the search process. Whether unsupervised representation learning can help QAS without any predictor is still an open topic. In this work, we propose a framework QAS with unsupervised representation learning and visualize how unsupervised architecture representation learning encourages quantum circuit architectures with similar connections and operators to cluster together. Specifically, our framework enables the process of QAS to be decoupled from unsupervised architecture representation learning so that the learned representation can be directly applied to different downstream applications. Furthermore, our framework is predictor-free eliminating the need for a large number of labeled quantum circuits.

One of the common hazards and issues in meteorology and agriculture is the problem of frost, chilling or freezing. This event occurs when the minimum ambient temperature falls below a certain value. This phenomenon causes a lot of damage to the country, especially Fars province. Solving this problem requires that, in addition to predicting the minimum temperature, we can provide enough time to implement the necessary measures. Empirical methods have been provided by the Food and Agriculture Organization (FAO), which can predict the minimum temperature, but not in time. In addition to this, we can use machine learning methods to model the minimum temperature. In this study, we have used three methods Gated Recurrent Unit (GRU), Temporal Convolutional Network (TCN) as deep learning methods, and Gradient Boosting (XGBoost). A customized loss function designed for methods based on deep learning, which can be effective in reducing prediction errors. With methods based on deep learning models, not only do we observe a reduction in RMSE error compared to empirical methods but also have more time to predict minimum temperature. Thus, we can model the minimum temperature for the next 24 hours by having the current 24 hours. With the gradient boosting model (XGBoost) we can keep the prediction time as deep learning and RMSE error reduced. Finally, we experimentally concluded that machine learning methods work better than empirical methods and XGBoost model can have better performance in this problem among other implemented.

Optimal transport (OT) is attracting increasing attention in machine learning. It aims to transport a source distribution to a target one at minimal cost. In its vanilla form, the source and target distributions are predetermined, which contracts to the real-world case involving undetermined targets. In this paper, we propose Doubly Bounded Optimal Transport (DB-OT), which assumes that the target distribution is restricted within two boundaries instead of a fixed one, thus giving more freedom for the transport to find solutions. Based on the entropic regularization of DB-OT, three scaling-based algorithms are devised for calculating the optimal solution. We also show that our DB-OT is helpful for barycenter-based clustering, which can avoid the excessive concentration of samples in a single cluster. Then we further develop DB-OT techniques for long-tailed classification which is an emerging and open problem. We first propose a connection between OT and classification, that is, in the classification task, training involves optimizing the Inverse OT to learn the representations, while testing involves optimizing the OT for predictions. With this OT perspective, we first apply DB-OT to improve the loss, and the Balanced Softmax is shown as a special case. Then we apply DB-OT for inference in the testing process. Even with vanilla Softmax trained features, our extensive experimental results show that our method can achieve good results with our improved inference scheme in the testing stage.

This document offers a critical overview of the emerging trends and significant advancements in artificial intelligence (AI) within the pharmaceutical industry. Detailing its application across key operational areas, including research and development, animal testing, clinical trials, hospital clinical stages, production, regulatory affairs, quality control and other supporting areas, the paper categorically examines AI's role in each sector. Special emphasis is placed on cutting-edge AI technologies like machine learning algorithms and their contributions to various aspects of pharmaceutical operations. Through this comprehensive analysis, the paper highlights the transformative potential of AI in reshaping the pharmaceutical industry's future.

The task of deducing three-dimensional molecular configurations from their two-dimensional graph representations holds paramount importance in the fields of computational chemistry and pharmaceutical development. The rapid advancement of machine learning, particularly within the domain of deep generative networks, has revolutionized the precision of predictive modeling in this context. Traditional approaches often adopt a two-step strategy: initially estimating interatomic distances and subsequently refining the spatial molecular structure by solving a distance geometry problem. However, this sequential approach occasionally falls short in accurately capturing the intricacies of local atomic arrangements, thereby compromising the fidelity of the resulting structural models. Addressing these limitations, this research introduces a cutting-edge generative framework named \method{}. This framework is grounded in the principles of diffusion observed in classical non-equilibrium thermodynamics. \method{} views atoms as discrete entities and excels in guiding the reversal of diffusion, transforming a distribution of stochastic noise back into coherent molecular structures through a process akin to a Markov chain. This transformation commences with the initial representation of a molecular graph in an abstract latent space, culminating in the realization of three-dimensional structures via a sophisticated bilevel optimization scheme meticulously tailored to meet the specific requirements of the task. One of the formidable challenges in this modeling endeavor involves preserving roto-translational invariance to ensure that the generated molecular conformations adhere to the laws of physics. Extensive experimental evaluations confirm the efficacy of the proposed \method{} in comparison to state-of-the-art methods.