Advances in the expressivity of pretrained models have increased interest in the design of adaptation protocols which enable safe and effective transfer learning. Going beyond conventional linear probing (LP) and fine tuning (FT) strategies, protocols that can effectively control feature distortion, i.e., the failure to update features orthogonal to the in-distribution, have been found to achieve improved outof-distribution generalization (OOD). In order to limit this distortion, the LP+FT protocol, which first learns a linear probe and then uses this initialization for subsequent FT, was proposed. However, in this paper, we find when adaptation protocols (LP, FT, LP+FT) are also evaluated on a variety of safety objectives (e.g., calibration, robustness, etc.), a complementary perspective to feature distortion is helpful to explain protocol behavior. To this end, we study the susceptibility of protocols to simplicity bias (SB), i.e. the well-known propensity of deep neural networks to rely upon simple features, as SB has recently been shown to underlie several problems in robust generalization. Using a synthetic dataset, we demonstrate the susceptibility of existing protocols to SB. Given the strong effectiveness of LP+FT, we then propose modified linear probes that help mitigate SB, and lead to better initializations for subsequent FT. We verify the effectiveness of the proposed LP+FT variants for decreasing SB in a controlled setting, and their ability to improve OOD generalization and safety on three adaptation datasets. Indeed, representations from such high-quality SSL models have been found to be more robust (Hendrycks et al., 2019; Liu et al., 2021), transferable (Ericsson et al., 2021) and semantically consistent (Caron et al., 2021) than their supervised counterparts. In this regard, there is growing need for adaptation protocols that explicitly capitalize on these improved pretrained features to induce similar beneficial properties, e.g., Figure 1: Strong and Safe Adaptation. Recently, however, Kumar et al. (2022) proved that by modifying features only in the ID representation subspace, FT can lead to higher OOD error as it distorts directions outside the ID subspace that are needed for OOD generalization. As both ID and OOD subspaces are represented by the pretrained model, Kumar et al. demonstrate that limiting feature distortion, or controlling updates towards the ID subspace, can lead to improved ID and OOD performance.

Modern deep neural networks are highly over-parameterized compared to the data on which they are trained, yet they often generalize remarkably well. A flurry of recent work has asked: why do deep networks not overfit to their training data? In this work, we make a series of empirical observations that investigate and extend the hypothesis that deeper networks are inductively biased to find solutions with lower effective rank embeddings. We conjecture that this bias exists because the volume of functions that maps to low effective rank embedding increases with depth. We show empirically that our claim holds true on finite width linear and non-linear models on practical learning paradigms and show that on natural data, these are often the solutions that generalize well. We then show that the simplicity bias exists at both initialization and after training and is resilient to hyper-parameters and learning methods. We further demonstrate how linear over-parameterization of deep non-linear models can be used to induce low-rank bias, improving generalization performance on CIFAR and ImageNet without changing the modeling capacity.

The vulnerability of machine learning models to adversarial attacks has been attracting considerable attention in recent years. Most existing studies focus on the behavior of stand-alone single-agent learners. In comparison, this work studies adversarial training over graphs, where individual agents are subjected to perturbations of varied strength levels across space. It is expected that interactions by linked agents, and the heterogeneity of the attack models that are possible over the graph, can help enhance robustness in view of the coordination power of the group. Using a min-max formulation of diffusion learning, we develop a decentralized adversarial training framework for multi-agent systems. We analyze the convergence properties of the proposed scheme for both convex and non-convex environments, and illustrate the enhanced robustness to adversarial attacks.

Empirical studies of the loss landscape of deep networks have revealed that many local minima are connected through low-loss valleys. Yet, little is known about the theoretical origin of such valleys. We present a general framework for finding continuous symmetries in the parameter space, which carve out low-loss valleys. Our framework uses equivariances of the activation functions and can be applied to different layer architectures. To generalize this framework to nonlinear neural networks, we introduce a novel set of nonlinear, data-dependent symmetries. These symmetries can transform a trained model such that it performs similarly on new samples, which allows ensemble building that improves robustness under certain adversarial attacks. We then show that conserved quantities associated with linear symmetries can be used to define coordinates along low-loss valleys. The conserved quantities help reveal that using common initialization methods, gradient flow only explores a small part of the global minimum. By relating conserved quantities to convergence rate and sharpness of the minimum, we provide insights on how initialization impacts convergence and generalizability.

Botnet attacks are a major threat to networked systems because of their ability to turn the network nodes that they compromise into additional attackers, leading to the spread of high volume attacks over long periods. The detection of such Botnets is complicated by the fact that multiple network IP addresses will be simultaneously compromised, so that Collective Classification of compromised nodes, in addition to the already available traditional methods that focus on individual nodes, can be useful. Thus this work introduces a collective Botnet attack classification technique that operates on traffic from an n-node IP network with a novel Associated Random Neural Network (ARNN) that identifies the nodes which are compromised. The ARNN is a recurrent architecture that incorporates two mutually associated, interconnected and architecturally identical n-neuron random neural networks, that act simultneously as mutual critics to reach the decision regarding which of n nodes have been compromised. A novel gradient learning descent algorithm is presented for the ARNN, and is shown to operate effectively both with conventional off-line training from prior data, and with on-line incremental training without prior off-line learning. Real data from a 107 node packet network is used with over 700,000 packets to evaluate the ARNN, showing that it provides accurate predictions. Comparisons with other well-known state of the art methods using the same learning and testing datasets, show that the ARNN offers significantly better performance.

This paper demonstrates that some non-classical models of human decision-making can be run successfully as circuits on quantum computers. Since the 1960s, many observed cognitive behaviors have been shown to violate rules based on classical probability and set theory. For example, the order in which questions are posed in a survey affects whether participants answer 'yes' or 'no', so the population that answers 'yes' to both questions cannot be modeled as the intersection of two fixed sets. It can, however, be modeled as a sequence of projections carried out in different orders. This and other examples have been described successfully using quantum probability, which relies on comparing angles between subspaces rather than volumes between subsets. Now in the early 2020s, quantum computers have reached the point where some of these quantum cognitive models can be implemented and investigated on quantum hardware, by representing the mental states in qubit registers, and the cognitive operations and decisions using different gates and measurements. This paper develops such quantum circuit representations for quantum cognitive models, focusing particularly on modeling order effects and decision-making under uncertainty. The claim is not that the human brain uses qubits and quantum circuits explicitly (just like the use of Boolean set theory does not require the brain to be using classical bits), but that the mathematics shared between quantum cognition and quantum computing motivates the exploration of quantum computers for cognition modeling. Key quantum properties include superposition, entanglement, and collapse, as these mathematical elements provide a common language between cognitive models, quantum hardware, and circuit implementations.

Mordecai3 is a new end-to-end text geoparser and event geolocation system. The system performs toponym resolution using a new neural ranking model to resolve a place name extracted from a document to its entry in the Geonames gazetteer. It also performs event geocoding, the process of linking events reported in text with the place names where they are reported to occur, using an off-the-shelf question-answering model. The toponym resolution model is trained on a diverse set of existing training data, along with several thousand newly annotated examples. The paper describes the model, its training process, and performance comparisons with existing geoparsers. The system is available as an open source Python library, Mordecai 3, and replaces an earlier geoparser, Mordecai v2, one of the most widely used text geoparsers (Halterman 2017).

We develop the use of mutual information (MI), a well-established metric in information theory, to interpret the inner workings of deep learning models. To accurately estimate MI from a finite number of samples, we present GMM-MI (pronounced $``$Jimmie$"$), an algorithm based on Gaussian mixture models that can be applied to both discrete and continuous settings. GMM-MI is computationally efficient, robust to the choice of hyperparameters and provides the uncertainty on the MI estimate due to the finite sample size. We extensively validate GMM-MI on toy data for which the ground truth MI is known, comparing its performance against established mutual information estimators. We then demonstrate the use of our MI estimator in the context of representation learning, working with synthetic data and physical datasets describing highly non-linear processes. We train deep learning models to encode high-dimensional data within a meaningful compressed (latent) representation, and use GMM-MI to quantify both the level of disentanglement between the latent variables, and their association with relevant physical quantities, thus unlocking the interpretability of the latent representation. We make GMM-MI publicly available.

Joint radar-communications (JRC) has emerged as a promising technology for efficiently using the limited electromagnetic spectrum. In JRC applications such as secure military receivers, often the radar and communications signals are overlaid in the received signal. In these passive listening outposts, the signals and channels of both radar and communications are unknown to the receiver. The ill-posed problem of recovering all signal and channel parameters from the overlaid signal is terms as dual-blind deconvolution (DBD). In this work, we investigate a more challenging version of DBD with a multi-antenna receiver. We model the radar and communications channels with a few (sparse) continuous-valued parameters such as time delays, Doppler velocities, and directions-of-arrival (DoAs). To solve this highly ill-posed DBD, we propose to minimize the sum of multivariate atomic norms (SoMAN) that depends on the unknown parameters. To this end, we devise an exact semidefinite program using theories of positive hyperoctant trigonometric polynomials (PhTP). Our theoretical analyses show that the minimum number of samples and antennas required for perfect recovery is logarithmically dependent on the maximum of the number of radar targets and communications paths rather than their sum. We show that our approach is easily generalized to include several practical issues such as gain/phase errors and additive noise. Numerical experiments show the exact parameter recovery for different JRC

Check out what's clicking on Foxnews.com. Russia stepped up its missile and drone attacks against Ukraine on Wednesday, killing students and other civilians, in a violent follow-up to dueling high-level diplomatic missions aimed at bringing peace after 13 months of war. "Russia is shelling the city with bestial savagery," President Volodymyr Zelenskyy wrote in a Telegram post accompanying video showing what he said was a Russian missile striking a nine-story apartment building on a busy road in the southeastern city of Zaporizhzhia. "Residential areas where ordinary people and children live are being fired at." At least one person was killed in the attack shown in the Zaporizhzhia video, apparently recorded by closed circuit TV cameras.