Gustavo Batista

Ayyoob Hamza, Hassan Habibi Gharakheili, Theophilus A. Benson et al.

IoT networks are increasingly becoming target of sophisticated new cyber-attacks. Anomaly-based detection methods are promising in finding new attacks, but there are certain practical challenges like false-positive alarms, hard to explain, and difficult to scale cost-effectively. The IETF recent standard called Manufacturer Usage Description (MUD) seems promising to limit the attack surface on IoT devices by formally specifying their intended network behavior. In this paper, we use SDN to enforce and monitor the expected behaviors of each IoT device, and train one-class classifier models to detect volumetric attacks. Our specific contributions are fourfold. (1) We develop a multi-level inferencing model to dynamically detect anomalous patterns in network activity of MUD-compliant traffic flows via SDN telemetry, followed by packet inspection of anomalous flows. This provides enhanced fine-grained visibility into distributed and direct attacks, allowing us to precisely isolate volumetric attacks with microflow (5-tuple) resolution. (2) We collect traffic traces (benign and a variety of volumetric attacks) from network behavior of IoT devices in our lab, generate labeled datasets, and make them available to the public. (3) We prototype a full working system (modules are released as open-source), demonstrates its efficacy in detecting volumetric attacks on several consumer IoT devices with high accuracy while maintaining low false positives, and provides insights into cost and performance of our system. (4) We demonstrate how our models scale in environments with a large number of connected IoTs (with datasets collected from a network of IP cameras in our university campus) by considering various training strategies (per device unit versus per device type), and balancing the accuracy of prediction against the cost of models in terms of size and training time.

Bo Chen, Ali Bakhshi, Gustavo Batista et al.

An increasing number of artificial intelligence (AI) applications involve the execution of deep neural networks (DNNs) on edge devices. Many practical reasons motivate the need to update the DNN model on the edge device post-deployment, such as refining the model, concept drift, or outright change in the learning task. In this paper, we consider the scenario where retraining can be done on the server side based on a copy of the DNN model, with only the necessary data transmitted to the edge to update the deployed model. However, due to bandwidth constraints, we want to minimise the transmission required to achieve the update. We develop a simple approach based on matrix factorisation to compress the model update -- this differs from compressing the model itself. The key idea is to preserve existing knowledge in the current model and optimise only small additional parameters for the update which can be used to reconstitute the model on the edge. We compared our method to similar techniques used in federated learning; our method usually requires less than half of the update size of existing methods to achieve the same accuracy.

Arman Pashamokhtari, Gustavo Batista, Hassan Habibi Gharakheili

Machine Learning-based techniques have shown success in cyber intelligence. However, they are increasingly becoming targets of sophisticated data-driven adversarial attacks resulting in misprediction, eroding their ability to detect threats on network devices. In this paper, we present AdIoTack, a system that highlights vulnerabilities of decision trees against adversarial attacks, helping cybersecurity teams quantify and refine the resilience of their trained models for monitoring IoT networks. To assess the model for the worst-case scenario, AdIoTack performs white-box adversarial learning to launch successful volumetric attacks that decision tree ensemble models cannot flag. Our first contribution is to develop a white-box algorithm that takes a trained decision tree ensemble model and the profile of an intended network-based attack on a victim class as inputs. It then automatically generates recipes that specify certain packets on top of the indented attack packets (less than 15% overhead) that together can bypass the inference model unnoticed. We ensure that the generated attack instances are feasible for launching on IP networks and effective in their volumetric impact. Our second contribution develops a method to monitor the network behavior of connected devices actively, inject adversarial traffic (when feasible) on behalf of a victim IoT device, and successfully launch the intended attack. Our third contribution prototypes AdIoTack and validates its efficacy on a testbed consisting of a handful of real IoT devices monitored by a trained inference model. We demonstrate how the model detects all non-adversarial volumetric attacks on IoT devices while missing many adversarial ones. The fourth contribution develops systematic methods for applying patches to trained decision tree ensemble models, improving their resilience against adversarial volumetric attacks.

Arman Pashamokhtari, Norihiro Okui, Masataka Nakahara et al.

Millions of vulnerable consumer IoT devices in home networks are the enabler for cyber crimes putting user privacy and Internet security at risk. Internet service providers (ISPs) are best poised to play key roles in mitigating risks by automatically inferring active IoT devices per household and notifying users of vulnerable ones. Developing a scalable inference method that can perform robustly across thousands of home networks is a non-trivial task. This paper focuses on the challenges of developing and applying data-driven inference models when labeled data of device behaviors is limited and the distribution of data changes (concept drift) across time and space domains. Our contributions are three-fold: (1) We collect and analyze network traffic of 24 types of consumer IoT devices from 12 real homes over six weeks to highlight the challenge of temporal and spatial concept drifts in network behavior of IoT devices; (2) We analyze the performance of two inference strategies, namely "global inference" (a model trained on a combined set of all labeled data from training homes) and "contextualized inference" (several models each trained on the labeled data from a training home) in the presence of concept drifts; and (3) To manage concept drifts, we develop a method that dynamically applies the ``closest'' model (from a set) to network traffic of unseen homes during the testing phase, yielding better performance in 20% of scenarios.

Arunan Sivanathan, David Warren, Deepak Mishra et al.

Machine learning models have demonstrated strong performance in classifying network traffic and identifying Internet-of-Things (IoT) devices, enabling operators to discover and manage IoT assets at scale. However, many existing approaches rely on end-to-end supervised pipelines or task-specific fine-tuning, resulting in traffic representations that are tightly coupled to labeled datasets and deployment environments, which can limit generalizability. In this paper, we study the problem of learning generalizable traffic representations for IoT device identification. We design compact encoder architectures that learn per-flow embeddings from unlabeled IoT traffic and evaluate them using a frozen-encoder protocol with a simple supervised classifier. Our specific contributions are threefold. (1) We develop unsupervised encoder--decoder models that learn compact traffic representations from unlabeled IoT network flows and assess their quality through reconstruction-based analysis. (2) We show that these learned representations can be used effectively for IoT device-type classification using simple, lightweight classifiers trained on frozen embeddings. (3) We provide a systematic benchmarking study against the state-of-the-art pretrained traffic encoders, showing that larger models do not necessarily yield more robust representations for IoT traffic. Using more than 18 million real IoT traffic flows collected across multiple years and deployment environments, we learn traffic representations from unlabeled data and evaluate device-type classification on disjoint labeled subsets, achieving macro F1-scores exceeding 0.9 for device-type classification and demonstrating robustness under cross-environment deployment.

Yunrui Zhang, Gustavo Batista, Salil S. Kanhere

Deep neural networks often produce miscalibrated probability estimates, leading to overconfident predictions. A common approach for calibration is fitting a post-hoc calibration map on unseen validation data that transforms predicted probabilities. A key desirable property of the calibration map is instance-wise monotonicity (i.e., preserving the ranking of probability outputs). However, most existing post-hoc calibration methods do not guarantee monotonicity. Previous monotonic approaches either use an under-parameterized calibration map with limited expressive ability or rely on black-box neural networks, which lack interpretability and robustness. In this paper, we propose a family of novel monotonic post-hoc calibration methods, which employs a constrained calibration map parameterized linearly with respect to the number of classes. Our proposed approach ensures expressiveness, robustness, and interpretability while preserving the relative ordering of the probability output by formulating the proposed calibration map as a constrained optimization problem. Our proposed methods achieve state-of-the-art performance across datasets with different deep neural network models, outperforming existing calibration methods while being data and computation-efficient. Our code is available at https://github.com/YunruiZhang/Calibration-by-Constrained-Transformation

Yunrui Zhang, Gustavo Batista, Salil S. Kanhere

Time series classification is usually regarded as a distinct task from tabular data classification due to the importance of temporal information. However, in this paper, by performing permutation tests that disrupt temporal information on the UCR time series classification archive, the most widely used benchmark for time series classification, we identify a significant proportion of datasets where temporal information has little to no impact on classification. Many of these datasets are tabular in nature or rely mainly on tabular features, leading to potentially biased evaluations of time series classifiers focused on temporal information. To address this, we propose UCR Augmented, a benchmark based on the UCR time series classification archive designed to evaluate classifiers' ability to extract and utilize temporal information. Testing classifiers from seven categories on this benchmark revealed notable shifts in performance rankings. Some previously overlooked approaches perform well, while others see their performance decline significantly when temporal information is crucial. UCR Augmented provides a more robust framework for assessing time series classifiers, ensuring fairer evaluations. Our code is available at https://github.com/YunruiZhang/Revisit-Time-Series-Classification-Benchmark.

Rushi Jayeshkumar Babaria, Minzhao Lyu, Gustavo Batista et al.

Network traffic classification is of great importance for network operators in their daily routines, such as analyzing the usage patterns of multimedia applications and optimizing network configurations. Internet service providers (ISPs) that operate high-speed links expect network flow classifiers to accurately classify flows early, using the minimal number of necessary initial packets per flow. These classifiers must also be robust to packet sequence disorders in candidate flows and capable of detecting unseen flow types that are not within the existing classification scope, which are not well achieved by existing methods. In this paper, we develop FastFlow, a time-series flow classification method that accurately classifies network flows as one of the known types or the unknown type, which dynamically selects the minimal number of packets to balance accuracy and efficiency. Toward the objectives, we first develop a flow representation process that converts packet streams at both per-packet and per-slot granularity for precise packet statistics with robustness to packet sequence disorders. Second, we develop a sequential decision-based classification model that leverages LSTM architecture trained with reinforcement learning. Our model makes dynamic decisions on the minimal number of time-series data points per flow for the confident classification as one of the known flow types or an unknown one. We evaluated our method on public datasets and demonstrated its superior performance in early and accurate flow classification. Deployment insights on the classification of over 22.9 million flows across seven application types and 33 content providers in a campus network over one week are discussed, showing that FastFlow requires an average of only 8.37 packets and 0.5 seconds to classify the application type of a flow with over 91% accuracy and over 96% accuracy for the content providers.

Bodu Gong, Gustavo Enrique Batista, Pierre Lafaye de Micheaux

In the era of large-scale neural network models, optimization algorithms often struggle with generalization due to an overreliance on training loss. One key insight widely accepted in the machine learning community is the idea that wide basins (regions around a local minimum where the loss increases gradually) promote better generalization by offering greater stability to small changes in input data or model parameters. In contrast, sharp minima are typically more sensitive and less stable. Motivated by two key empirical observations - the inherent heavy-tailed distribution of gradient noise in stochastic gradient descent and the Edge of Stability phenomenon during neural network training, in which curvature grows before settling at a plateau, we introduce Adaptive Heavy Tailed Stochastic Gradient Descent (AHTSGD). The algorithm injects heavier-tailed noise into the optimizer during the early stages of training to enhance exploration and gradually transitions to lighter-tailed noise as sharpness stabilizes. By dynamically adapting to the sharpness of the loss landscape throughout training, AHTSGD promotes accelerated convergence to wide basins. AHTSGD is the first algorithm to adjust the nature of injected noise into an optimizer based on the Edge of Stability phenomenon. AHTSGD consistently outperforms SGD and other noise-based methods on benchmarks like MNIST and CIFAR-10, with marked gains on noisy datasets such as SVHN. It ultimately accelerates early training from poor initializations and improves generalization across clean and noisy settings, remaining robust to learning rate choices.

Zeinab Ebrahimi, Gustavo Batista, Mohammad Deghat

Stochastic gradient descent (SGD) and its many variants are the widespread optimization algorithms for training deep neural networks. However, SGD suffers from inevitable drawbacks, including vanishing gradients, lack of theoretical guarantees, and substantial sensitivity to input. The Alternating Direction Method of Multipliers (ADMM) has been proposed to address these shortcomings as an effective alternative to the gradient-based methods. It has been successfully employed for training deep neural networks. However, ADMM-based optimizers have a slow convergence rate. This paper proposes an Anderson Acceleration for Deep Learning ADMM (AA-DLADMM) algorithm to tackle this drawback. The main intention of the AA-DLADMM algorithm is to employ Anderson acceleration to ADMM by considering it as a fixed-point iteration and attaining a nearly quadratic convergence rate. We verify the effectiveness and efficiency of the proposed AA-DLADMM algorithm by conducting extensive experiments on four benchmark datasets contrary to other state-of-the-art optimizers.

Yanping Chen, Adena Why, Gustavo Batista et al.

The ability to use inexpensive, noninvasive sensors to accurately classify flying insects would have significant implications for entomological research, and allow for the development of many useful applications in vector control for both medical and agricultural entomology. Given this, the last sixty years have seen many research efforts on this task. To date, however, none of this research has had a lasting impact. In this work, we explain this lack of progress. We attribute the stagnation on this problem to several factors, including the use of acoustic sensing devices, the over-reliance on the single feature of wingbeat frequency, and the attempts to learn complex models with relatively little data. In contrast, we show that pseudo-acoustic optical sensors can produce vastly superior data, that we can exploit additional features, both intrinsic and extrinsic to the insect's flight behavior, and that a Bayesian classification approach allows us to efficiently learn classification models that are very robust to over-fitting. We demonstrate our findings with large scale experiments that dwarf all previous works combined, as measured by the number of insects and the number of species considered.