Dr. Achraf Cohen

Modern foundational models are often compressed via a combination of structured pruning and re-training to meet the strict compute, memory, and connectivity constraints of edge deployments. While state-of-the-art (SoTA) pruning schemes target the entire Transformer, we adopt a simple, layer-wise L 2 -norm pruning on only the multi-layer perceptron (MLP) blocks as a fixed baseline. Our focus is not on achieving maximal compression, but on isolating the impact of the re-training loss function: (i) L2-norm Pruning with Cross-Entropy Fine-Tuning (L2PFT), which relies on labeled data, versus (ii) L2-norm Pruning with KL-Divergence Self-Distillation (L2PSD), which utilizes only teacher logits without requiring labeled data. We evaluate both pipelines on the OLMo2-7B-SFT model for CommonsenseQA, suitable for intermittent or denied connectivity scenarios typical of edge networks. Under identical pruning schedules, L2PSD achieves comparable or superior test accuracy to L2PFT, indicating that the choice of loss function has a significant impact on compressed model recovery in resource-constrained environments.

Transfer learning is commonly utilized in various fields such as computer vision, natural language processing, and medical imaging due to its impressive capability to address sub-tasks and work with different datasets. However, its application in cybersecurity has not been thoroughly explored. In this paper, we present an innovative neurosymbolic AI framework designed for network intrusion detection systems, which play a crucial role in combating malicious activities in cybersecurity. Our framework leverages transfer learning and uncertainty quantification. The findings indicate that transfer learning models, trained on large and well-structured datasets, outperform neural-based models that rely on smaller datasets, paving the way for a new era in cybersecurity solutions.

Academic advising plays a crucial role in fostering student success. However, challenges such as limited advisor availability can hinder effective support. Generative AI, particularly AI-powered chatbots, offers the potential to enhance student advising in higher education by providing personalized guidance. These technologies help college students find the information and resources needed to create degree plans aligned with their academic goals. This research introduces ARGObot, an intelligent advising system that facilitates student navigation of university policies through automated interpretation of the student handbook as its primary knowledge base. ARGObot enhances accessibility to critical academic policies and procedures, supporting incoming students' success through personalized guidance. Our system integrates a multifunctional agent enhanced by a Large Language Model (LLM). The architecture employs multiple external tools to enhance its capabilities: a Retrieval-Augmented Generation (RAG) system accesses verified university sources; email integration facilitates Human-in-the-Loop (HITL) interaction; and a web search function expands the system's knowledge base beyond predefined constraints. This approach enables the system to provide contextually relevant and verified responses to various student queries. This architecture evolved from our initial implementation based on Gemini 1 Pro, which revealed significant limitations due to its lack of agent-based functionality, resulting in hallucination issues and irrelevant responses. Subsequent evaluation demonstrated that our enhanced version, integrating GPT-4 with the text-embedding-ada-002 model, achieved superior performance across all metrics. This paper also presents a comparative analysis of both implementations, highlighting the architectural improvements and their impact on system performance.

IoT systems such as smart homes or cities capture large amounts of data in real time using a wide range of interconnected sensor devices linked to edge nodes and their associated cloud services. In such systems, sensors must operate continuously and reliably in their deployed environment to deliver meaningful data to the users. As time passes, sensors may fail or become less effective in collecting raw data, resulting in erroneous data being delivered to the end user, who relies on their accuracy for decision-making. In this paper, we study detection methods of anomalies in sensor data based on Hotelling's T^{2} statistic. In simulation experiments, we calculate the lowest possible shift in data streams that are detectable by our proposed methods using sensor data from a smart home. Our results show that the proposed method successfully identifies the low shift in the means, given a confidence level with an accuracy of 90%. These results are based on data and do not assume a specific distribution, which opens up new potential applications for IoT data and anomaly detection. Additionally, some low shifts can be difficult to detect visually, but our proposed method will be able to detect those low shifts.

This research presents economic operation of distributed energy resources in an islanded microgrid considering uncertainties associated with forecasting error of load, wind, and solar series. The forecasting is performed using artificial neural networks (ANN). A distribution of the forecasting errors was fitted. A discrete set of probabilities was used to create a set of possible scenarios representing possible deviations from the forecasted outputs of load, wind, and solar. The problem of economical operation of the electric grid was formulated as a stochastic optimization model to minimize expected total cost (ETC). The ETC expression consists of (a) expected operating cost of the generators, which includes linearized fuel cost and startup costs (b) expected operating costs of energy storage system and (c) expected interruption cost. Load demand data from New England area was tested to study the efficacy of this approach. A test system consisting of 10 conventional generators, 100 wind turbines at a maximum of 0.14 MW each, totaling 14 MW and 100 solar panels at a maximum of 0.36 MW each, totaling 36 MW was considered. Simulations were carried out in MATLAB and the results show that coordinating all energy resources (1) significantly enhances the power management capability of the grid (2) reduces the ETC in addition to (3) maintaining grid balance under high penetration of renewable energy sources.

In this paper, we present a result regarding the probability distribution of wavelets coefficients. It extends the Theorem 1 that is presented in our previous work -Cohen, A. et al. Design of experiments and statistical process control using wavelets analysis. Control Engineering Practice. 2015- to include Biorthogonal wavelets. Then, a new control chart, called OWave (Orthogonal Wavelets), is proposed in order to detect mean change in the process. The statistic of the proposed control chart is based on weighted wavelets coefficients. Performance study shows that OWave control chart performs slightly better than the optimal version EWMA control chart.

Autocorrelation and non-normality of process characteristic variables are two main difficulties that industrial engineers must face when they should implement control charting techniques. This paper presents new issues regarding the probability distribution of wavelets coefficients. Firstly, we highlight that wavelets coefficients have capacities to strongly decrease autocorrelation degree of original data and are normally-like distributed, especially in the case of Haar wavelet. We used AR(1) model with positive autoregressive parameters to simulate
autocorrelated data. Illustrative examples are presented to show wavelets coefficients properties. Secondly, the distributional parameters of wavelets coefficients are derived, it shows that wavelets coefficients reflect an interesting statistical properties for SPC purposes.