Mohammadreza Doostmohammadian’s Post

Thrilled to announce that my research paper "Identifying Spam Accounts on Instagram: An Analysis of User Activity Data Using Machine Learning" is now officially available on IEEE Xplore! You can access the paper here: https://lnkd.in/gX3rQVQq I'd love to hear your thoughts and discuss how this work can further be explored😊 #Research #MachineLearning #SocialMedia

Our paper in IEEE Transactions on Circuits and Systems I on "veriSIMPLER: An Automated Formal Verification Methodology for SIMPLER MAGIC Design Style Based In-Memory Computing" is now available #online as #EarlyAccess https://lnkd.in/emsSC_Rx Universität Bremen Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI) Data Science Center (DSC) der Universität Bremen Chandan Kumar Jha Khushboo Qayyum Kemal Çağlar Coşkun Simranjeet Singh Muhammad Hassan Rainer Leupers Farhad Merchant #LogicInMemory #formal #verification #magic

How can we solve complex decision-making problems under uncertainties to achieve the best possible outcomes? At the 62nd IEEE Conference on Decision and Control (CDC 2023), our researchers proposed an algorithm to solve complex decision-making problems over a specific timeframe. This involves making choices that impact multiple factors (MDP problems), by building a bilinear program to make the best possible decisions. To Know More- bit.ly/3V8cPe3 Authors- Uday Kumar, Veeraruna Kavitha, Sanjay P. Bhat, Nandyala Hemachandra #Rersearch #Process #Algorithm

I'm excited to share our latest blog post, "Understanding SIMD: Infinite Complexity of Trivial Problems." In this insightful article, we delve into the intricacies of Single Instruction, Multiple Data (SIMD) and explore how seemingly simple problems can become infinitely complex when approached from this perspective. Whether you're a seasoned developer or just curious about advanced computing concepts, this post offers valuable insights into the applications and implications of SIMD in today's technology landscape. Read the full article here: https://ift.tt/bJQUFKa

Article Alert!   Title: A Graph Learning-Based Approach for Lateral Movement Detection Authors: Mahdi Rabbani; Leila Rashidi; Ali A. Ghorbani Journal: IEEE Transactions on Network and Service Management   Link: https://lnkd.in/gqspWgdN   Abstract: Lateral movement, a crucial phase in the Advanced Persistent Threat (APT) life cycle, refers to a strategy employed by adversaries to traverse horizontally within a network. The aim is to gain access to various systems or resources, thereby expanding their control and potential access to valuable targets. Detecting these attacks becomes challenging for conventional detection systems due to various factors, including the complexity of pathways, the mimicking of legitimate user behavior by attackers, and limited network visibility. To address these challenges, advanced detection techniques are required to effectively and dynamically analyze multiple features within the interconnected structure of the network. This paper introduces an innovative approach to detect malicious lateral movement paths by leveraging authentication events and graph learning techniques. The proposed method involves constructing a heterogeneous graph, and employing DeepWalk for node embedding. By combining node embedding features with the temporal information of authentication events, feature vectors are generated for each authentication request. These features are then used to train multiple machine learning-based classifiers to detect malicious lateral movement paths. Furthermore, to assess the model’s performance in a more realistic scenario, a series of additional experiments were conducted. These experiments provided further validation of the model’s robustness and its capability for forward prediction. #GraphLearning #MachineLearning #LateralMovementDetection #AdvancedPersistentThreat

In computer science, Data Structures are ways of organizing and storing data in a computer so that it can be accessed and modified efficiently. They define the layout and arrangement of data in memory, enabling algorithms to manipulate the data easily.

Can the performance of Bayesian optimizers be increased by computing performance metrics after the modelling process? That is the question we sort to answer as part of the Bayesian Optimization Hackathon for Chemistry and Materials (https://lnkd.in/gyAqDA-w). Turns out that it can (and can't). Computing metrics after the modelling process can result in significant computational savings if multiple metrics can be derived from the same models, however there wasn't a noticeable improvement in the number of experiments that are required before optima convergence. Post-modelling metric computation often gets an early lead due to the biases post-modelling computation introduces, but the lead is typically lost due to the biases working against it in later iterations. It's possible that these limitations could be overcome by implementing non-constant priors in the Bayesian process. The biases may also prove useful in domain reduction applications due to their tendency to explore regions with the highest potential based on metric bounds. You can read more about this work on the GitHub page: https://lnkd.in/g6eGgHkp. Big thanks to Sterling G. Baird for making the event and happen and to Acceleration Consortium and Merck KGaA, Darmstadt, Germany for sponsoring it.

Book chapter about C-Idea introduction, a fast algorithm method for computing emerging closed data cubes,  IGI Global (January 2019). My research is available on @ResearchGate: https://lnkd.in/eGJiRc-V

My research paper, "Behavioral Characteristics-Based Drowsiness Detection System Using Machine Learning," has been published in an IEEE conference 🎉 This study focuses on developing a system that leverages machine learning to detect drowsiness based on behavioral cues, aiming to improve safety by alerting users before accidents occur. You can explore the full paper here: https://lnkd.in/gvM9pQV5 Thanks to everyone who supported and contributed to this work 🙏 #research #machinelearning #IEEE #drowsinessdetection #safetytech

⚡ Seeking faster inference speeds without sacrificing accuracy? To answer this question, I recently tried out the latest 💡QoQ (quattuor-octo-quattuor), a W4A8KV4 quantization algorithm with 4-bit weight, 8-bit activation, and 4-bit KV cache on 🔍Llama-3-8B-Instruct-262k. First step was to generate QoQ quantized checkpoints using LMQuant and dump the fake-quantized models. Afterwards, Qserve provides a checkpoint converter to real-quantize and pack the model into QServe format I ran the throughput benchmark on 1x A100 in order to compare the findings with the Qserve documented values for Llama-3-8B on A100. 📈 Impressive results achieved! With an average throughput of 2925 tok/s over 3 rounds and a batch size of 256, QoQ showcases its efficiency and scalability. 🤗 Huggingface: https://lnkd.in/dsmd5qxq ⚙️Qserve: https://lnkd.in/duHyQx7U ⚙️lmquant: https://lnkd.in/dq4XhDMM