Glioblastoma Multiforme (GBM) is an aggressive and highly heterogeneous brain tumor with poor survival outcomes. While conventional radiomic analyses focus on tumor-centric regions, emerging surgical strategies, such as GTR and supratotal resection (SupTR), highlight the importance of the peritumoral zone. Moreover, due to intratumoral heterogeneity, MGMTpm, a key molecular biomarker guiding chemotherapeutic decisions, is assessed invasively and remains prone to sampling bias. Therefore, this study examined the prognostic and predictive utility of deep radiomic features from tumor and peritumoral regions in preoperative MRI to improve OS prediction and enable non-invasive MGMTpm classification. Multi-parametric structural MRI scans (T1, T1-Gd, T2, and T2-FLAIR) from 520 to 200 GBM patients were analyzed for OS and MGMTpm prediction, respectively. Tumor and peritumoral masks were segmented and expanded using morphological dilation from 2 to 12 mm. 11,000 deep features per patient were extracted using ResNet50 and ViT-B16 models, capturing patterns that may reflect tumor infiltration and microenvironmental changes. Hybrid feature selection using variance thresholding and Recursive Feature Elimination (RFE) was applied, including age and gender. Support Vector Machine (SVM) classifiers were trained using 10-fold cross-validation. In OS prediction, the inclusion of an 8 mm peritumoral margin improved AUC from 0.74 (95
Amplified Reflection Distributed Denial of Service (AR-DDoS) attack has become a serious attack on the SDN controller, and the control channel bandwidth degrades network performance. In an AR-DDoS attack, attackers exploit vulnerable reflectors and amplify traffic towards the victim, overwhelming network resources. This saturates network bandwidth, the centralized controller, and the flow tables, and increases queuing delay. Existing solutions detect AR-DDoS attacks using sketches and traffic statistics experience suboptimal detection accuracy due to inadequately designed algorithms. Moreover, previous solutions demonstrate machine learning accuracy in classifying attack traffic; they face difficulty in detecting emerging AR-DDoS attacks. The ever-evolving attack landscape, containing new features, necessitates continuous model retraining. In this paper, we propose a resource-efficient Quantized Convolutional Neural Network (QCNN) model deployed within the programmable data plane, designed for intelligent traffic monitoring and accurate detection of AR-DDoS attacks. Specifically, quantization of a Convolution Neural Network model is performed to easily infer the model in a resource-constrained programmable data plane. Further, this model is mapped to the match and action tables to efficiently detect and mitigate attack traffic. Furthermore, a self-learning mechanism is used to improve detection in a dynamic network environment. In addition, in-network traffic prioritization has been performed with PQ-Sketch to prevent queue congestion and drop attack flows. Extensive experiments on the Tofino hardware testbed show that QCNN achieves 99.83, 98.66, and 98.95
Large Language Models (LLMs) have demonstrated strong capabilities as autonomous agents through tool use, planning, and decision-making abilities, leading to their widespread adoption across diverse tasks. As task complexity grows, multi-agent LLM systems are increasingly used to solve problems collaboratively. However, safety and security of these systems remains largely under-explored. Existing benchmarks and datasets predominantly focus on single-agent settings, failing to capture the unique vulnerabilities of multi-agent dynamics and co-ordination. To address this gap, we introduce $\textbf{T}$hreats and $\textbf{A}$ttacks in $\textbf{M}$ulti-$\textbf{A}$gent $\textbf{S}$ystems ($\textbf{TAMAS}$), a benchmark designed to evaluate the robustness and safety of multi-agent LLM systems. TAMAS includes five distinct scenarios comprising 300 adversarial instances across six attack types and 211 tools, along with 100 harmless tasks. We assess system performance across ten backbone LLMs and three agent interaction configurations from Autogen and CrewAI frameworks, highlighting critical challenges and failure modes in current multi-agent deployments. Furthermore, we introduce Effective Robustness Score (ERS) to assess the tradeoff between safety and task effectiveness of these frameworks. Our findings show that multi-agent systems are highly vulnerable to adversarial attacks, underscoring the urgent need for stronger defenses. TAMAS provides a foundation for systematically studying and improving the safety of multi-agent LLM systems.
We design new minimal-subpacketization schemes for information-theoretic private information retrieval on graph-based replicated databases. In graph-based replication, the system consists of K files replicated across N servers according to a graph with N vertices and K edges. The client wants to retrieve one desired file, while keeping the index of the desired file private from each server via a query-response protocol. We seek PIR protocols that have (a) high rate, which is the ratio of the file-size to the total download cost, and (b) low subpacketization, which acts as a constraint on the size of the files for executing the protocol. We report two new schemes which have unit-subpacketization (which is minimal): (i) for a special class of graphs known as star graphs, and (ii) for general graphs. Our star-graph scheme has a better rate than previously known schemes with low subpacketization for general star graphs. Our scheme for general graphs uses a decomposition of the graph via independent sets. This scheme achieves a rate lower than prior schemes for the complete graph, however it can achieve higher rates than known for some specific graph classes. An extension of our scheme to the case of multigraphs achieves a higher rate than previous schemes for the complete multi-graph.
Question Answering (QA) tasks, which involve extracting answers from a given context, are relatively straightforward for modern Large Language Models (LLMs) when the context is short. However, long contexts pose challenges due to the quadratic complexity of the self-attention mechanism. This challenge is compounded in Indic languages, which are often low-resource. This study explores context-shortening techniques, including Open Information Extraction (OIE), coreference resolution, Answer Paragraph Selection (APS), and their combinations, to improve QA performance. Compared to the baseline of unshortened (long) contexts, our experiments on four Indic languages (Hindi, Tamil, Telugu, and Urdu) demonstrate that context-shortening techniques yield an average improvement of 4 https://github.com/ritwikmishra/IndicGenQA .
