A5112988008- Network Security and Intrusion Detection
- Advanced Malware Detection Techniques
- Topic Modeling
- Nanowire Synthesis and Applications
- Ga2O3 and related materials
- Security and Verification in Computing
- ZnO doping and properties
- Perovskite Materials and Applications
- Natural Language Processing Techniques
- Digital and Cyber Forensics
- Quantum Dots Synthesis And Properties
- Gas Sensing Nanomaterials and Sensors
- GaN-based semiconductor devices and materials
- Advanced Sensor and Energy Harvesting Materials
- Electronic and Structural Properties of Oxides
- Advanced MEMS and NEMS Technologies
- Artificial Intelligence in Healthcare and Education
- Integrated Circuits and Semiconductor Failure Analysis
- Advanced Database Systems and Queries
- Quantum and electron transport phenomena
- Digital Media Forensic Detection
- Semiconductor Quantum Structures and Devices
- Radiomics and Machine Learning in Medical Imaging
- Semantic Web and Ontologies
- Model-Driven Software Engineering Techniques
Indian Institute of Technology Guwahati
2025
University of New Orleans
2015-2020
A high-performance broad band UV/visible photodetector has been successfully fabricated on a fully wide bandgap ZnO/ZnS type-II heterojunction core/shell nanowire array. The device can detect photons with energies significantly smaller (2.2 eV) than the gap of ZnO (3.2 and ZnS (3.7 eV), which is mainly attributed to spatially indirect transition facilitated by abrupt interface between core shell. performance was further enhanced through piezo-phototronic effect induced lowering barrier...
We release Code Llama, a family of large language models for code based on Llama 2 providing state-of-the-art performance among open models, infilling capabilities, support input contexts, and zero-shot instruction following ability programming tasks. provide multiple flavors to cover wide range applications: foundation (Code Llama), Python specializations - Python), instruction-following Instruct) with 7B, 13B, 34B 70B parameters each. All are trained sequences 16k tokens show improvements...
To advance the photodetection capabilities of type II core/shell nanowire‐based photodetectors, ZnSe has been employed to coat on ZnO nanowire array creating a type‐II heterojunction with additional lower indirect band gap (≈1.84 eV), via facile two‐step process chemical vapor deposition and pulse laser ablation. The integrated photodetector based ZnO/ZnSe structure is capable detecting entire range visible spectrum, from blue red excitation source, as well UV light. absolute sensitivity...
In the above article, an incorrect version of Figure 5 was used in which polarity parts (c) and (d) reversed.The correct v ersion 5c,d is shown accompanying this Correction.This means that a sentence final paragraph Section 2.3 should be amended read as follows:"When nanowires are bent under external load, appearance negative piezopotential at compressive surface reduces hole trapping barriers (Figure 5c) Ag-ZnTe CdSe/ZnTe interfaces eliminates electron trap states CdSe/ ZnTe interface,...
This paper presents CyberSecEval, a comprehensive benchmark developed to help bolster the cybersecurity of Large Language Models (LLMs) employed as coding assistants. As what we believe be most extensive unified safety date, CyberSecEval provides thorough evaluation LLMs in two crucial security domains: their propensity generate insecure code and level compliance when asked assist cyberattacks. Through case study involving seven models from Llama 2, Code Llama, OpenAI GPT large language...
Large language models (LLMs) introduce new security risks, but there are few comprehensive evaluation suites to measure and reduce these risks. We present BenchmarkName, a novel benchmark quantify LLM risks capabilities. two areas for testing: prompt injection code interpreter abuse. evaluated multiple state-of-the-art (SOTA) LLMs, including GPT-4, Mistral, Meta Llama 3 70B-Instruct, Code Llama. Our results show that conditioning away risk of attack remains an unsolved problem; example, all...
Historically, hands-on cybersecurity exercises helped reinforce the basic concepts. However, most of them focused on user level attacks and defenses did not provide a convenient way studying kernel security. Since OS kernels foundations for applications, any compromise to will lead computer that cannot be trusted. Moreover, there has been great interest in using virtualization profile, characterize, observe events including security incidents. Virtual Machine Introspection (VMI) is technique...
File fragment classification is an essential problem in digital forensics. Although several attempts had been made to solve this challenging problem, a general solution has not found. In work, we propose hierarchical machine-learning-based approach with optimized support vector machines (SVM) as the base classifiers for file classification. This consists of more at top level and specialized fine-grain lower levels hierarchy. We also primitive taxonomy types that can be used perform evaluate...
Piezoelectric-related multiproperty coupling effects have recently attracted much attention for developing next generation multifunctional devices. In this paper, the among magnetic field, piezoelectricity, and photoexcitation were investigated in ZnO nanowire arrays. The response of current flowing through array to applied field decreased with increasing strength. However, due piezo-magnetotronic effect, magnetically induced was magnified by 1 order magnitude applying an external stress....
Identification of operating system kernel version is essential in a large number forensic and security applications both cloud local environments. Prior state-of-the-art uses complex differential analysis several aspects implementation knowledge data structures. In this paper, we present working research prototype codeid-elf for ELF binaries based on its Windows counterpart codeid, which can identify kernels through relocation entries extracted from the binaries. We show that...
We are releasing a new suite of security benchmarks for LLMs, CYBERSECEVAL 3, to continue the conversation on empirically measuring LLM cybersecurity risks and capabilities. 3 assesses 8 different across two broad categories: risk third parties, application developers end users. Compared previous work, we add areas focused offensive capabilities: automated social engineering, scaling manual cyber operations, autonomous operations. In this paper discuss applying these Llama models...
W. Zhou and co-workers demonstrate a prototype piezo-phototronic photodetector integrated on 3D CdSe/ZnTe core/shell nanowire array. As they discuss in article number 1400050, the broad band photo-detection can be achieved with two orders of magnitude change absolute current along four responsivity, which demonstrates enhanced effect group II–VI heterojunction.
The piezo‐phototronic effect is of immense importance for improving the performance optoelectronic nanodevices. This accomplished by tuning charge carrier generation, separation, and transport under influence inner piezopotential. In this paper, a broad band photodetector demonstrated that based on II‐VI binary CdSe/ZnTe core/shell nanowire arrays, in which photodetection greatly enhanced effect. UV (385 nm), blue (465 green (520 nm) illumination infers saturation free response with an...