To date, there has been no systematic investigation of thermal profiles of keyboards, and thus no efforts have been made to secure them. This serves as our main motivation for constructing a means for password harvesting from keyboard thermal emanations. Specifically, we introduce Thermanator: a new post-factum insider attack based on heat transfer caused by a user typing a password on a typical external (plastic) keyboard. We conduct and describe a user study that collected thermal residues from 30 users entering 10 unique passwords (both weak and strong) on 4 popular commodity keyboards. Results show that entire sets of key-presses can be recovered by non-expert users as late as 30 seconds after initial password entry, while partial sets can be recovered as late as 1 minute after entry. However, the thermal residue side-channel lacks information about password length, duplicate key-presses, and key-press ordering. To overcome these limitations, we leverage keyboard acoustic emanations and combine the two to yield AcuTherm, the first hybrid side-channel attack on keyboards. AcuTherm significantly reduces password search without the need for any training on the victim's typing. We report results gathered for many representative passwords based on a user study involving 19 subjects. The takeaway of this work is three-fold: (1) using plastic keyboards to enter secrets (such as passwords and PINs) is even less secure than previously recognized, (2) post-factum thermal imaging attacks are realistic, and (3) hybrid (multiple side-channel) attacks are both realistic and effective.

Data streams are often defined as large amounts of data flowing continuously at high speed. Moreover, these data are likely subject to changes in data distribution, known as concept drift. Given all the reasons mentioned above, learning from streams is often online and under restrictions of memory consumption and run-time. Although many classification algorithms exist, most of the works published in the area use Naive Bayes (NB) and Hoeffding Trees (HT) as base learners in their experiments. This article proposes an in-depth evaluation of k-Nearest Neighbors (k-NN) as a candidate for classifying data streams subjected to concept drift. It also analyses the complexity in time and the two main parameters of k-NN, i.e., the number of nearest neighbors used for predictions (k), and window size (w). We compare different parameter values for k-NN and contrast it to NB and HT both with and without a drift detector (RDDM) in many datasets. We formulated and answered 10 research questions which led to the conclusion that k-NN is a worthy candidate for data stream classification, especially when the run-time constraint is not too restrictive.

Despite the constant evolution of similarity searching research, it continues to face the same challenges stemming from the complexity of the data, such as the curse of dimensionality and computationally expensive distance functions. Various machine learning techniques have proven capable of replacing elaborate mathematical models with combinations of simple linear functions, often gaining speed and simplicity at the cost of formal guarantees of accuracy and correctness of querying. The authors explore the potential of this research trend by presenting a lightweight solution for the complex problem of 3D protein structure search. The solution consists of three steps -- (i) transformation of 3D protein structural information into very compact vectors, (ii) use of a probabilistic model to group these vectors and respond to queries by returning a given number of similar objects, and (iii) a final filtering step which applies basic vector distance functions to refine the result.

We study the problem of episodic reinforcement learning in continuous state-action spaces with unknown rewards and transitions. Specifically, we consider the setting where the rewards and transitions are modeled using parametric bilinear exponential families. We propose an algorithm, BEF-RLSVI, that a) uses penalized maximum likelihood estimators to learn the unknown parameters, b) injects a calibrated Gaussian noise in the parameter of rewards to ensure exploration, and c) leverages linearity of the exponential family with respect to an underlying RKHS to perform tractable planning. We further provide a frequentist regret analysis of BEF-RLSVI that yields an upper bound of $\tilde{\mathcal{O}}(\sqrt{d^3H^3K})$, where $d$ is the dimension of the parameters, $H$ is the episode length, and $K$ is the number of episodes. Our analysis improves the existing bounds for the bilinear exponential family of MDPs by $\sqrt{H}$ and removes the handcrafted clipping deployed in existing \RLSVI-type algorithms. Our regret bound is order-optimal with respect to $H$ and $K$.

Active learning promises to improve annotation efficiency by iteratively selecting the most important data to be annotated first. However, we uncover a striking contradiction to this promise: active learning fails to select data as efficiently as random selection at the first few choices. We identify this as the cold start problem in vision active learning, caused by a biased and outlier initial query. This paper seeks to address the cold start problem by exploiting the three advantages of contrastive learning: (1) no annotation is required; (2) label diversity is ensured by pseudo-labels to mitigate bias; (3) typical data is determined by contrastive features to reduce outliers. Experiments are conducted on CIFAR-10-LT and three medical imaging datasets (i.e. Colon Pathology, Abdominal CT, and Blood Cell Microscope). Our initial query not only significantly outperforms existing active querying strategies but also surpasses random selection by a large margin. We foresee our solution to the cold start problem as a simple yet strong baseline to choose the initial query for vision active learning. Code is available: https://github.com/c-liangyu/CSVAL

We consider a multi-armed bandit problem with $M$ latent contexts, where an agent interacts with the environment for an episode of $H$ time steps. Depending on the length of the episode, the learner may not be able to estimate accurately the latent context. The resulting partial observation of the environment makes the learning task significantly more challenging. Without any additional structural assumptions, existing techniques to tackle partially observed settings imply the decision maker can learn a near-optimal policy with $O(A)^H$ episodes, but do not promise more. In this work, we show that learning with {\em polynomial} samples in $A$ is possible. We achieve this by using techniques from experiment design. Then, through a method-of-moments approach, we design a procedure that provably learns a near-optimal policy with $O(\texttt{poly}(A) + \texttt{poly}(M,H)^{\min(M,H)})$ interactions. In practice, we show that we can formulate the moment-matching via maximum likelihood estimation. In our experiments, this significantly outperforms the worst-case guarantees, as well as existing practical methods.

In the last two decades we are witnessing a huge increase of valuable big data structured in the form of graphs or networks. To apply traditional machine learning and data analytic techniques to such data it is necessary to transform graphs into vector-based representations that preserve the most essential structural properties of graphs. For this purpose, a large number of graph embedding methods have been proposed in the literature. Most of them produce general-purpose embeddings suitable for a variety of applications such as node clustering, node classification, graph visualisation and link prediction. In this paper, we propose two novel graph embedding algorithms based on random walks that are specifically designed for the node classification problem. Random walk sampling strategies of the proposed algorithms have been designed to pay special attention to hubs -- high-degree nodes that have the most critical role for the overall connectedness in large-scale graphs. The proposed methods are experimentally evaluated by analyzing the classification performance of three classification algorithms trained on embeddings of real-world networks. The obtained results indicate that our methods considerably improve the predictive power of examined classifiers compared to currently the most popular random walk method for generating general-purpose graph embeddings (node2vec).

Extracting relevant urban patterns from multiple data sources can be difficult using classical clustering algorithms since we have to make a suitable setup of the hyperparameters of the algorithms and deal with outliers. It should be addressed correctly to help urban planners in the decision-making process for the further development of a big city. For instance, experts' main interest in criminology is comprehending the relationship between crimes and the socio-economic characteristics at specific georeferenced locations. In addition, the classical clustering algorithms take little notice of the intricate spatial correlations in georeferenced data sources. This paper presents a new approach to detecting the most relevant urban patterns from multiple data sources based on tensor decomposition. Compared to classical methods, the proposed approach's performance is attested to validate the identified patterns' quality. The result indicates that the approach can effectively identify functional patterns to characterize the data set for further analysis in achieving good clustering quality. Furthermore, we developed a generic framework named TensorAnalyzer, where the effectiveness and usefulness of the proposed methodology are tested by a set of experiments and a real-world case study showing the relationship between the crime events around schools and students performance and other variables involved in the analysis.

In this paper, we consider privacy aspects of wireless federated learning (FL) with Over-the-Air (OtA) transmission of gradient updates from multiple users/agents to an edge server. By exploiting the waveform superposition property of multiple access channels, OtA FL enables the users to transmit their updates simultaneously with linear processing techniques, which improves resource efficiency. However, this setting is vulnerable to privacy leakage since an adversary node can hear directly the uncoded message. Traditional perturbation-based methods provide privacy protection while sacrificing the training accuracy due to the reduced signal-to-noise ratio. In this work, we aim at minimizing privacy leakage to the adversary and the degradation of model accuracy at the edge server at the same time. More explicitly, spatially correlated perturbations are added to the gradient vectors at the users before transmission. Using the zero-sum property of the correlated perturbations, the side effect of the added perturbation on the aggregated gradients at the edge server can be minimized. In the meanwhile, the added perturbation will not be canceled out at the adversary, which prevents privacy leakage. Theoretical analysis of the perturbation covariance matrix, differential privacy, and model convergence is provided, based on which an optimization problem is formulated to jointly design the covariance matrix and the power scaling factor to balance between privacy protection and convergence performance. Simulation results validate the correlated perturbation approach can provide strong defense ability while guaranteeing high learning accuracy.

Embeddings, which compress information in raw text into semantics-preserving low-dimensional vectors, have been widely adopted for their efficacy. However, recent research has shown that embeddings can potentially leak private information about sensitive attributes of the text, and in some cases, can be inverted to recover the original input text. To address these growing privacy challenges, we propose a privatization mechanism for embeddings based on homomorphic encryption, to prevent potential leakage of any piece of information in the process of text classification. In particular, our method performs text classification on the encryption of embeddings from state-of-the-art models like BERT, supported by an efficient GPU implementation of CKKS encryption scheme. We show that our method offers encrypted protection of BERT embeddings, while largely preserving their utility on downstream text classification tasks.