Many applications of text generation require incorporating different constraints to control the semantics or style of generated text. These constraints can be hard (e.g., ensuring certain keywords are included in the output) and soft (e.g., contextualizing the output with the left- or right-hand context). In this paper, we present Energy-based Constrained Decoding with Langevin Dynamics (COLD), a decoding framework which unifies constrained generation as specifying constraints through an energy function, then performing efficient differentiable reasoning over the constraints through gradient-based sampling. COLD decoding is a flexible framework that can be applied directly to off-the-shelf left-to-right language models without the need for any task-specific fine-tuning, as demonstrated through three challenging text generation applications: lexically-constrained generation, abductive reasoning, and counterfactual reasoning. Our experiments on these constrained generation tasks point to the effectiveness of our approach, both in terms of automatic and human evaluation.

Federated Recommendation (FR) has received considerable popularity and attention in the past few years. In FR, for each user, its feature vector and interaction data are kept locally on its own client thus are private to others. Without the access to above information, most existing poisoning attacks against recommender systems or federated learning lose validity. Benifiting from this characteristic, FR is commonly considered fairly secured. However, we argue that there is still possible and necessary security improvement could be made in FR. To prove our opinion, in this paper we present FedRecAttack, a model poisoning attack to FR aiming to raise the exposure ratio of target items. In most recommendation scenarios, apart from private user-item interactions (e.g., clicks, watches and purchases), some interactions are public (e.g., likes, follows and comments). Motivated by this point, in FedRecAttack we make use of the public interactions to approximate users' feature vectors, thereby attacker can generate poisoned gradients accordingly and control malicious users to upload the poisoned gradients in a well-designed way. To evaluate the effectiveness and side effects of FedRecAttack, we conduct extensive experiments on three real-world datasets of different sizes from two completely different scenarios. Experimental results demonstrate that our proposed FedRecAttack achieves the state-of-the-art effectiveness while its side effects are negligible. Moreover, even with small proportion (3%) of malicious users and small proportion (1%) of public interactions, FedRecAttack remains highly effective, which reveals that FR is more vulnerable to attack than people commonly considered.

The implicit biases of gradient-based optimization algorithms are conjectured to be a major factor in the success of modern deep learning. In this work, we investigate the implicit bias of gradient flow and gradient descent in two-layer fully-connected neural networks with leaky ReLU activations when the training data are nearly-orthogonal, a common property of high-dimensional data. For gradient flow, we leverage recent work on the implicit bias for homogeneous neural networks to show that asymptotically, gradient flow produces a neural network with rank at most two. Moreover, this network is an $\ell_2$-max-margin solution (in parameter space), and has a linear decision boundary that corresponds to an approximate-max-margin linear predictor. For gradient descent, provided the random initialization variance is small enough, we show that a single step of gradient descent suffices to drastically reduce the rank of the network, and that the rank remains small throughout training. We provide experiments which suggest that a small initialization scale is important for finding low-rank neural networks with gradient descent.

In recent years, we have seen a handful of work on inference algorithms over non-stationary data streams. Given their flexibility, Bayesian non-parametric models are a good candidate for these scenarios. However, reliable streaming inference under the concept drift phenomenon is still an open problem for these models. In this work, we propose a variational inference algorithm for Dirichlet process mixture models. Our proposal deals with the concept drift by including an exponential forgetting over the prior global parameters. Our algorithm allows to adapt the learned model to the concept drifts automatically. We perform experiments in both synthetic and real data, showing that the proposed model is competitive with the state-of-the-art algorithms in the density estimation problem, and it outperforms them in the clustering problem.

We introduce in this paper a new statistical perspective, exploiting the Jaccard similarity metric, as a measure-based metric to effectively invoke non-linear features in the loss of self-supervised contrastive learning. Specifically, our proposed metric may be interpreted as a dependence measure between two adapted projections learned from the so-called latent representations. This is in contrast to the cosine similarity measure in the conventional contrastive learning model, which accounts for correlation information. To the best of our knowledge, this effectively non-linearly fused information embedded in the Jaccard similarity, is novel to self-supervision learning with promising results. The proposed approach is compared to two state-of-the-art self-supervised contrastive learning methods on three image datasets. We not only demonstrate its amenable applicability in current ML problems, but also its improved performance and training efficiency.

A successful paradigm in representation learning is to perform self-supervised pretraining using tasks based on mini-batch statistics (e.g., SimCLR, VICReg, SwAV, MSN). We show that in the formulation of all these methods is an overlooked prior to learn features that enable uniform clustering of the data. While this prior has led to remarkably semantic representations when pretraining on class-balanced data, such as ImageNet, we demonstrate that it can hamper performance when pretraining on class-imbalanced data. By moving away from conventional uniformity priors and instead preferring power-law distributed feature clusters, we show that one can improve the quality of the learned representations on real-world class-imbalanced datasets. To demonstrate this, we develop an extension of the Masked Siamese Networks (MSN) method to support the use of arbitrary features priors.

Massive multiple input multiple output (MIMO) systems are typically designed under the assumption of linear power amplifiers (PAs). However, PAs are typically most energy-efficient when operating close to their saturation point, where they cause non-linear distortion. Moreover, when using conventional precoders, this distortion coherently combines at the user locations, limiting performance. As such, when designing an energy-efficient massive MIMO system, this distortion has to be managed. In this work, we propose the use of a neural network (NN) to learn the mapping between the channel matrix and the precoding matrix, which maximizes the sum rate in the presence of this non-linear distortion. This is done for a third-order polynomial PA model for both the single and multi-user case. By learning this mapping a significant increase in energy efficiency is achieved as compared to conventional precoders and even as compared to perfect digital pre-distortion (DPD), in the saturation regime.

On a SAS discussion forum, a statistical programmer asked about how to understand the statistics that are displayed when you use the TEST statement in PROC REG (or other SAS regression procedures) to test for linear relationships between regression coefficients. The documentation for the TEST statement in PROC REG explains the F test in terms of a matrix of linear constraints. However, the programmer wanted a simpler explanation. Fortunately, there is an easy way to explain the TEST statement from first principles. The explanation involves running two regression models.

Quantization refers to a technique where we express a range of values by a single quantum value. For images, this means that we can compress an entire color range into one specific color. This technique is lossy i.e. we intentionally lose information in favor of lower memory consumption. In this tutorial, I will show you how to implement color quantization yourself with very few lines of code. We are going to use Python with scikit-learn, numpy, PIL, and matplotlib. Lets's start by downloading the beautiful image of "Old Man of Storr" taken by Pascal van Soest which we will work with (if you are on Windows or do not have access to wget simply download the image and save it as image.jpeg):

The field of robotics and machine intelligence has experienced significant growth over the last few years. Technological advancements have enabled researchers to develop increasingly sophisticated devices and artificial intelligence systems capable of performing tasks traditionally reserved for humans. To understand where these fields are heading, it's important to review where they've been and what the role of robotics and machine intelligence has been so far in human history. Robots are machines that can complete physical tasks autonomously or semi-autonomously without human intervention. Robots typically perform tasks for humans, such as manufacturing, assembly, welding, painting, and many other functions. They may be guided remotely or operate autonomously using artificial intelligence-based systems such as GPS or vision sensors where they locate objects through image recognition technology. The exact definition of a robot is open to interpretation. Machines exhibit intelligence when they're able to sense, learn, and respond autonomously. But how is machine intelligence different from artificial intelligence?