We study stochastic optimization algorithms for constrained nonconvex stochastic optimization problems with Markovian data. In particular, we focus on the case when the transition kernel of the Markov chain is state-dependent. Such stochastic optimization problems arise in various machine learning problems including strategic classification and reinforcement learning. For this problem, we study both projection-based and projection-free algorithms. In both cases, we establish that the number of calls to the stochastic first-order oracle to obtain an appropriately defined $\epsilon$-stationary point is of the order $\mathcal{O}(1/\epsilon^{2.5})$. In the projection-free setting we additionally establish that the number of calls to the linear minimization oracle is of order $\mathcal{O}(1/\epsilon^{5.5})$. We also empirically demonstrate the performance of our algorithm on the problem of strategic classification with neural networks.

Large pretrained language models generate fluent text but are notoriously hard to controllably sample from. In this work, we study constrained sampling from such language models: generating text that satisfies user-defined constraints, while maintaining fluency and the model's performance in a downstream task. We propose MuCoLa -- a sampling procedure that combines the log-likelihood of the language model with arbitrary (differentiable) constraints in a single energy function, and then generates samples in a non-autoregressive manner. Specifically, it initializes the entire output sequence with noise and follows a Markov chain defined by Langevin Dynamics using the gradients of the energy function. We evaluate MuCoLa on text generation with soft and hard constraints as well as their combinations obtaining significant improvements over competitive baselines for toxicity avoidance, sentiment control, and keyword-guided generation.

A matrix free and a low rank approximation preconditioner are proposed to accelerate the convergence of stochastic gradient descent (SGD) by exploiting curvature information sampled from Hessian-vector products or finite differences of parameters and gradients similar to the BFGS algorithm. Both preconditioners are fitted with an online updating manner minimizing a criterion that is free of line search and robust to stochastic gradient noise, and further constrained to be on certain connected Lie groups to preserve their corresponding symmetry or invariance, e.g., orientation of coordinates by the connected general linear group with positive determinants. The Lie group's equivariance property facilitates preconditioner fitting, and its invariance property saves any need of damping, which is common in second-order optimizers, but difficult to tune. The learning rate for parameter updating and step size for preconditioner fitting are naturally normalized, and their default values work well in most situations.

The second class of methods uses graph embedding, Many real-world systems can be represented as graphs where the in particular graph autoencoders. Graph convolutional network different entities are presented by nodes and their interactions by (GCN) based methods such as Graph autoencoders (GAE) [13], variational edges. An important task in studying large datasets is graph clustering. GAE (VGAE) [14], adversarially regularized graph autoencoder While there has been a lot of work on graph clustering using the (ARGA), adversarially regularized variational graph autoencoder connectivity between the nodes, many real-world networks also have (ARVGA) [15] and marginalized graph autoencoder for graph node attributes. Clustering attributed graphs requires joint modeling clustering (MGAE) [16] have demonstrated state-of-the-art performance of graph structure and node attributes. Recent work has focused on on several attributed graph clustering tasks. These methods graph convolutional networks and graph convolutional filters to combine directly use GCN as a feature extractor, where each convolutional structural and content information. However, these methods are layer is equivalent to a first-order low-pass graph filter that only takes mostly limited to lowpass filtering and do not explicitly optimize the the most immediate neighbors of each node into account.

Artificial neural networks are functions depending on a finite number of parameters typically encoded as weights and biases. The identification of the parameters of the network from finite samples of input-output pairs is often referred to as the \emph{teacher-student model}, and this model has represented a popular framework for understanding training and generalization. Even if the problem is NP-complete in the worst case, a rapidly growing literature -- after adding suitable distributional assumptions -- has established finite sample identification of two-layer networks with a number of neurons $m=\mathcal O(D)$, $D$ being the input dimension. For the range $D

In a spoofing attack, an attacker impersonates a legitimate user to access or tamper with data intended for or produced by the legitimate user. In wireless communication systems, these attacks may be detected by relying on features of the channel and transmitter radios. In this context, a popular approach is to exploit the dependence of the received signal strength (RSS) at multiple receivers or access points with respect to the spatial location of the transmitter. Existing schemes rely on long-term estimates, which makes it difficult to distinguish spoofing from movement of a legitimate user. This limitation is here addressed by means of a deep neural network that implicitly learns the distribution of pairs of short-term RSS vector estimates. The adopted network architecture imposes the invariance to permutations of the input (commutativity) that the decision problem exhibits. The merits of the proposed algorithm are corroborated on a data set that we collected.

The clinical interest is often to measure the volume of a structure, which is typically derived from a segmentation. In order to evaluate and compare segmentation methods, the similarity between a segmentation and a predefined ground truth is measured using popular discrete metrics, such as the Dice score. Recent segmentation methods use a differentiable surrogate metric, such as soft Dice, as part of the loss function during the learning phase. In this work, we first briefly describe how to derive volume estimates from a segmentation that is, potentially, inherently uncertain or ambiguous. This is followed by a theoretical analysis and an experimental validation linking the inherent uncertainty to common loss functions for training CNNs, namely cross-entropy and soft Dice. We find that, even though soft Dice optimization leads to an improved performance with respect to the Dice score and other measures, it may introduce a volume bias for tasks with high inherent uncertainty. These findings indicate some of the method's clinical limitations and suggest doing a closer ad-hoc volume analysis with an optional re-calibration step.

This article has been accepted at the IEEE Sensors 2022 Conference. Abstract--In this study, to address the current high earlydetection miss rate of colorectal cancer (CRC) polyps, we explore the potentials of utilizing transfer learning and machine learning (ML) classifiers to precisely and sensitively classify the type of CRC polyps. Instead of using the common colonoscopic images, we applied three different ML algorithms on the 3D textural image outputs of a unique vision-based surface tactile sensor (VS-TS). To collect realistic textural images of CRC polyps for training the utilized ML classifiers and evaluating their performance, we first designed and additively manufactured 48 types of realistic polyp phantoms with different hardness, type, and textures. Next, the performance of the used three ML algorithms in classifying the type of fabricated polyps was quantitatively evaluated using various statistical metrics.

Abstract: Currently, the engineering of miniature spectrometers mainly faces three problems: the mismatch between the number of filters at the front end of the detector and the spectral reconstruction accuracy; the lack of a stable spectral reconstruction algorithm; and the lack of a spectral reconstruction evaluation method suitable for engineering. Therefore, based on 20 sets of filters, this paper classifies and optimizes the filter array by the K-means algorithm and particle swarm algorithm, and obtains the optimal filter combination under different matrix dimensions. Then, the truncated singular value decomposition-convex optimization algorithm is used for high-precision spectral reconstruction, and the detailed spectral reconstruction process of two typical target spectra is described. In terms of spectral evaluation, due to the strong randomness of the target detected during the working process of the spectrometer, the standard value of the target spectrum cannot be obtained. Therefore, for the first time, we adopt the method of joint cross-validation of multiple sets of data for spectral evaluation.

I have rarely been as enthusiastic about a new research direction. We call them GFlowNets, for Generative Flow Networks. They live somewhere at the intersection of reinforcement learning, deep generative models and energy-based probabilistic modelling. They are also related to variational models and inference and I believe open new doors for non-parametric Bayesian modelling, generative active learning, and unsupervised or self-supervised learning of abstract representations to disentangle both the explanatory causal factors and the mechanisms that relate them. What I find exciting is that they open so many doors, but in particular for implementing the system 2 inductive biases I have been discussing in many of my papers and talks since 2017, that I argue are important to incorporate causality and deal with out-of-distribution generalization in a rational way.