The Information Bottleneck (IB) principle has emerged as a promising approach for enhancing the generalization, robustness, and interpretability of deep neural networks, demonstrating efficacy across image segmentation, document clustering, and semantic communication. Among IB implementations, the IB Lagrangian method, employing Lagrangian multipliers, is widely adopted. While numerous methods for the optimizations of IB Lagrangian based on variational bounds and neural estimators are feasible, their performance is highly dependent on the quality of their design, which is inherently prone to errors. To address this limitation, we introduce Structured IB, a framework for investigating potential structured features. By incorporating auxiliary encoders to extract missing informative features, we generate more informative representations. Our experiments demonstrate superior prediction accuracy and task-relevant information preservation compared to the original IB Lagrangian method, even with reduced network size.

The information bottleneck (IB) principle has been adopted to explain deep learning in terms of information compression and prediction, which are balanced by a trade-off hyperparameter. How to optimize the IB principle for better robustness and figure out the effects of compression through the trade-off hyperparameter are two challenging problems. Previous methods attempted to optimize the IB principle by introducing random noise into learning the representation and achieved state-of-the-art performance in the nuisance information compression and semantic information extraction. However, their performance on resisting adversarial perturbations is far less impressive. To this end, we propose an adversarial information bottleneck (AIB) method without any explicit assumptions about the underlying distribution of the representations, which can be optimized effectively by solving a Min-Max optimization problem. Numerical experiments on synthetic and real-world datasets demonstrate its effectiveness on learning more invariant representations and mitigating adversarial perturbations compared to several competing IB methods. In addition, we analyse the adversarial robustness of diverse IB methods contrasting with their IB curves, and reveal that IB models with the hyperparameter $\beta$ corresponding to the knee point in the IB curve achieve the best trade-off between compression and prediction, and has best robustness against various attacks.

The information bottleneck (IB) problem tackles the issue of obtaining relevant compressed representations T of some random variable X for the task of predicting Y. It is defined as a constrained optimization problem which maximizes the information the representation has about the task, I(T;Y), while ensuring that a minimum level of compression r is achieved; i.e., I(X;T) <= r. For practical reasons the problem is usually solved by maximizing the IB Lagrangian for many values of the Lagrange multiplier, therefore drawing the IB curve (i.e., the curve of maximal I(T;Y) for a given I(X;T)) and selecting the representation of desired predictability and compression. It is known when Y is a deterministic function of X, the IB curve cannot be explored, and other Lagrangians have been proposed to tackle this problem; e.g., the squared IB Lagrangian. In this paper, we (i) present a general family of Lagrangians which allow for the exploration of the IB curve in all scenarios; and (ii) prove that if these Lagrangians are used, there is a (and we know the) one-to-one mapping between the Lagrange multiplier and the desired compression rate r for known IB curve shapes, hence, freeing us from the burden of solving the optimization problem for many values of the Lagrange multiplier. That is, we can solve the original constrained problem with a single optimization.

Information bottleneck (IB) is a method for extracting information from one random variable X that is relevant for predicting another random variable Y . To do so, IB identifies an intermediate "bottleneck" variable T that has low mutual information I(X; T) and high mutual information I(Y; T). The IB curve characterizes the set of bottleneck variables that achieve maximal I(Y; T) for a given I(X; T), and is typically explored by optimizing the IB Lagrangian, I(Y; T) βI(X; T). Recently, there has been interest in applying IB to supervised learning, particularly for classification problems that use neural networks. In most classification problems, the output class Y is a deterministic function of the input X, which we refer to as "deterministic supervised learning". We demonstrate three pathologies that arise when IB is used in any scenario where Y is a deterministic function of X: (1) the IB curve cannot be recovered by optimizing the IB Lagrangian for different values of β; (2) there are "uninteresting" solutions at all points of the IB curve; and (3) for classifiers that achieve low error rates, the activity of different hidden layers will not exhibit a strict tradeoff between compression and prediction, contrary to a recent proposal. To address problem (1), we propose a functional that, unlike the IB Lagrangian, can recover the IB curve in all cases. We finish by demonstrating these issues on the MNIST dataset. The information bottleneck (IB) method (Tishby et al., 1999) provides a principled way to extract information that is present in one variable that is relevant for predicting another variable. Given two random variables X and Y, IB posits a "bottleneck" variable T that obeys the Markov condition Y X T . By the data processing inequality (DPI) (Cover & Thomas, 2012), this Markov condition implies that I(X; T) I(Y; T), meaning the bottleneck variable cannot contain more information about Y than it does about X.

We review the problem of defining and inferring a "state" for a control system based on complex, high-dimensional, highly uncertain measurement streams such as videos. Such a state, or representation, should contain all and only the information needed for control, and discount nuisance variability in the data. It should also have finite complexity, ideally modulated depending on available resources. This representation is what we want to store in memory in lieu of the data, as it "separates" the control task from the measurement process. For the trivial case with no dynamics, a representation can be inferred by minimizing the Information Bottleneck Lagrangian in a function class realized by deep neural networks. The resulting representation has much higher dimension than the data, already in the millions, but it is smaller in the sense of information content, retaining only what is needed for the task. This process also yields representations that are invariant to nuisance factors and having maximally independent components. We extend these ideas to the dynamic case, where the representation is the posterior density of the task variable given the measurements up to the current time, which is in general much simpler than the prediction density maintained by the classical Bayesian filter. Again this can be finitely-parametrized using a deep neural network, and already some applications are beginning to emerge. No explicit assumption of Markovianity is needed; instead, complexity trades off approximation of an optimal representation, including the degree of Markovianity.

Using established principles from Information Theory and Statistics, we show that in a deep neural network invariance to nuisance factors is equivalent to information minimality of the learned representation, and that stacking layers and injecting noise during training naturally bias the network towards learning invariant representations. We then show that, in order to avoid memorization, we need to limit the quantity of information stored in the weights, which leads to a novel usage of the Information Bottleneck Lagrangian on the weights as a learning criterion. This also has an alternative interpretation as minimizing a PAC-Bayesian bound on the test error. Finally, we exploit a duality between weights and activations induced by the architecture, to show that the information in the weights bounds the minimality and Total Correlation of the layers, therefore showing that regularizing the weights explicitly or implicitly, using SGD, not only helps avoid overfitting, but also fosters invariance and disentangling of the learned representation. The theory also enables predicting sharp phase transitions between underfitting and overfitting random labels at precise information values, and sheds light on the relation between the geometry of the loss function, in particular so-called "flat minima," and generalization.