Mohamed A Khamis | Egypt-Japan University of Science and Technology (E-JUST) (original) (raw)

Papers by Mohamed A Khamis

arXiv (Cornell University), Aug 23, 2016

EPJ Data Science

The aim of this paper is to analyze the Arab peoples reactions and attitudes towards the Russo-Uk... more The aim of this paper is to analyze the Arab peoples reactions and attitudes towards the Russo-Ukraine War through the social media of posted tweets, as a fast means to express opinions. We scrapped over 3 million tweets using some keywords that are related to the war and performed sentiment, emotion, and partiality analyses. For sentiment analysis, we employed a voting technique of several pre-trained Arabic language foundational models. For emotion analysis, we utilized a pre-constructed emotion lexicon. The partiality is analyzed through classifying tweets as being ‘Pro-Russia’, ‘Pro-Ukraine’, or ‘Neither’; and it indicates the bias or empathy towards either of the conflicting parties. This was achieved by constructing a weighted lexicon of n-grams related to either side. We found that the majority of the tweets carried ‘Negative’ sentiment. Emotions were not that obvious with a lot of tweets carrying ‘Mixed Feelings’. The more decisive tweets conveyed either ‘Joy’ or ‘Anger’ emo...

In this paper, we assess the scoring, ranking, docking, and screening powers of deep learning and... more In this paper, we assess the scoring, ranking, docking, and screening powers of deep learning and random forest scoring functions. For the scoring power, the DL RF scoring function (arithmetic mean between DL and RF scores) achieves Pearson's correlation coefficient between the predicted and experimentally measured binding affinities of 0.799 versus 0.758 of the RF scoring function. For the ranking power, the DL scoring function ranks the ligands bound to fixed target protein with accuracy 54% for the high-level ranking (correctly ranking the three ligands bound to the same target protein in a cluster) and with accuracy 78% for the low-level ranking (correctly ranking the best ligand only in the cluster) while the RF scoring function achieves (46% and 62%) respectively. For the docking power, the DL RF scoring function has a success rate when the three best-scored ligand binding poses are considered within 2Å root-mean-square-deviation from the native pose of 36.0% versus 30.2% of the RF scoring function. For the screening power, the DL scoring function has an average enrichment factor and success rate at the top 1% level of (2.69 and 6.45%) respectively versus (1.61 and 4.84%) respectively of the RF scoring function.

An in-Vehicle System and Method for During Accident Detection without being Fixed to Vehicle, 2020

An in-vehicle mobile system and method for during accident detection comprising acceleration, sou... more An in-vehicle mobile system and method for during accident detection comprising acceleration, sound sensors, communication devices, and doppler analysis component, in which the system is not fixed to the vehicle; in which a flying detection unit utilizes the acceleration sensor to detect the unfixed devices flying behaviour during a crash; in which the low sound frequency emitted from the motor is used to detect relative system’s speed to the motor; in which the mobile system can be placed anywhere inside the vehicle; in which doppler processing happens while the mobile device is flying before crashing into a surface; in which the accident severity level is detected and communicated to a centralised server.

System and Method for During Crash Accident Detection and Notification, 2020

A mobile system and method for accident detection comprising sensors for detecting acceleration a... more A mobile system and method for accident detection comprising sensors for detecting acceleration and orientation, communication device for communication with a central server, and processor for detecting the crash and severity of which during the occurrence for such an accident without relying on after crashing sensed data; in which acceleration sensors are of low-dynamic range; in which a component detects the change of vehicle speed using sampled acceleration and orientation data and corresponding frequency components of which; in which a component further computes the accident duration using said data; in which a fall detection components detects falls; in which accident severity component detects the accident severity using speed change, crash duration, and said fall information; in which the component sets the number of severity level with respect to the accelerometer dynamic range of the sensor; in which an accident direction components uses accelerometer and orientation sensors to find and report the crash direction and communicate through the communication device to a server in which notification is further forwarded to mobile devices.

2023 International Joint Conference on Neural Networks (IJCNN), 2023

Neural Computing and Applications

Human activity recognition (HAR) using inertial motion data has gained a lot of momentum in recen... more Human activity recognition (HAR) using inertial motion data has gained a lot of momentum in recent years both in research and industrial applications. From the abstract perspective, this has been driven by the rapid dynamics for building intelligent, smart environments, and ubiquitous systems that cover all aspects of human life including healthcare, sports, manufacturing, commerce, etc., which necessitate and subsume activity recognition aiming at recognizing the actions, characteristics, and goals of one or more agent(s) from a temporal series of observations streamed from one or more sensors. From a more concrete and seemingly orthogonal perspective, such momentum has been driven by the ubiquity of inertial motion sensors on-board mobile and wearable devices including smartphones, smartwatches, etc. In this paper we give an introductory and a comprehensive survey to the subject from a given perspective. We focus on a subset of topics, that we think are major, that will have signi...

Traffic fatalities are increasing in developing countries where there are few investments in road... more Traffic fatalities are increasing in developing countries where there are few investments in road safety. Culture and road conditions also affect driving habits. Therefore, automatic detection and reporting of driver behavior to concerned entities can potentially save lives. In particular, we analyze a driving maneuvers dataset collected from one environment (country) but tested in another environment with aggressive driving habits and irregular road conditions. We also develop an on-edge system with fast response time to serve users on a large scale. Specifically, we propose an approach for detecting aggressive and normal events using random forest classifier. We utilize the accelerometer and gyroscope smartphone readings to classify driving maneuvers events to five types (aggressive acceleration, suddenly break, aggressive turn right, aggressive turn left, and normal). We achieved an accuracy of only 63.4% by training our model on an available dataset collected from a foreigner environment and tested on our environment. The lowest precision value was 54% while the lowest recall was 42%. However, we achieved an accuracy of 98.4% when augmenting an available dataset with data collected with our application.
The lowest precision value was 98% while the lowest recall was 90%. From the results, it is shown that the available datasets do not generalize well to different driving habits and road conditions. Finally, an implementation of the random forest model using OpenCV on an Android platform is analyzed.

Accidents fatality is generally dependent on the time an emergency service is dispatched to the a... more Accidents fatality is generally dependent on the time an emergency service is dispatched to the accident scene. Decreasing this time requires fast and accurate accident detection and notification systems. Therefore, existing well-established systems rely on rugged devices, with specialised hardware and accurate sensors to allow for in-vehicle detection and notification. Smartphones have been considered as an alternative mainly due to their much lower cost. In this paper, we show that reliable accident detection can be achieved using main stream smartphone sensors (e.g., accelerometer and gyroscope). The method relies on detecting the accident pulse through using Fourier transform and a random forest classifier. The method also utilises a moving window to incorporate time; and is simple enough to allow for during-accident detection, not requiring the mobile to survive the accident. We have validated the model using the Ollie car-like robot micro accident-testbed and the gold standard in accident simulation, LS-DYNA, achieving a true positive rate of about 96% and true negative rate of 99%.

Several intelligent transportation systems focus on studying the various driver behaviors for num... more Several intelligent transportation systems focus on studying the various driver behaviors for numerous objectives. This includes the ability to analyze driver actions, sensitivity, distraction, and response time. As the data collection is one of the major concerns for learning and validating different driving situations, we present a driver behavior switching model validated by a low-cost data collection solution using smartphones. The proposed model is validated using a real dataset to predict the driver behavior in short duration periods. A literature survey on motion detection (specifically driving behavior detection using smartphones) is presented. Multiple Markov Switching Variable Auto-Regression (MSVAR) models are implemented to achieve a sophisticated fitting with the collected driver behavior data. This yields more accurate predictions not only for driver behavior but also for the entire driving situation. The performance of the presented models together with a suitable mod...

IEEE Geoscience and Remote Sensing Letters

ArXiv, 2021

Engineering Applications of Artificial Intelligence, 2015

Engineering Applications of Artificial Intelligence, 2014

Artificial Intelligence in Medicine, 2015

The objective of this paper is to highlight the state-of-the-art machine learning (ML) techniques... more The objective of this paper is to highlight the state-of-the-art machine learning (ML) techniques in computational docking. The use of smart computational methods in the life cycle of drug design is relatively a recent development that has gained much popularity and interest over the last few years. Central to this methodology is the notion of computational docking which is the process of predicting the best pose (orientation + conformation) of a small molecule (drug candidate) when bound to a target larger receptor molecule (protein) in order to form a stable complex molecule. In computational docking, a large number of binding poses are evaluated and ranked using a scoring function. The scoring function is a mathematical predictive model that produces a score that represents the binding free energy, and hence the stability, of the resulting complex molecule. Generally, such a function should produce a set of plausible ligands ranked according to their binding stability along with their binding poses. In more practical terms, an effective scoring function should produce promising drug candidates which can then be synthesized and physically screened using high throughput screening process. Therefore, the key to computer-aided drug design is the design of an efficient highly accurate scoring function (using ML techniques). The methods presented in this paper are specifically based on ML techniques. Despite many traditional techniques have been proposed, the performance was generally poor. Only in the last few years started the application of the ML technology in the design of scoring functions; and the results have been very promising. The ML-based techniques are based on various molecular features extracted from the abundance of protein-ligand information in the public molecular databases, e.g., protein data bank bind (PDBbind). In this paper, we present this paradigm shift elaborating on the main constituent elements of the ML approach to molecular docking along with the state-of-the-art research in this area. For instance, the best random forest (RF)-based scoring function [35] on PDBbind v2007 achieves a Pearson correlation coefficient between the predicted and experimentally determined binding affinities of 0.803 while the best conventional scoring function achieves 0.644 [34]. The best RF-based ranking power [6] ranks the ligands correctly based on their experimentally determined binding affinities with accuracy 62.5% and identifies the top binding ligand with accuracy 78.1%. We conclude with open questions and potential future research directions that can be pursued in smart computational docking; using molecular features of different nature (geometrical, energy terms, pharmacophore), advanced ML techniques (e.g., deep learning), combining more than one ML models.

The car following behavior has recently gained much attention due to its wide variety of applicat... more The car following behavior has recently gained much attention due to its wide variety of applications. This includes accident analysis, driver assessment, support systems, and road design. In this paper, we present a model that leverages Markov regime switching models to classify various car following regimes. The detected car following regimes are then mined to calibrate the parameters of drivers to be dependent on the driver’s current driving regime. A two stage Markov regime switching model is utilized to detect different car following regimes. The ﬁrst stage discriminates normal car following regimes from abnormal ones, while the second stage classiﬁes normal car following regimes to their ﬁne-grained regimes like braking, accelerating, standing, free-ﬂowing, and normal following. A genetic algorithm is then employed to the observed driver data in each car following regime to optimize car following model parameter values of the driver in each regime. Experimental evaluation of the proposed model using a real dataset shows that it can detect up-normal (rare and short time) events. In addition, it can infer the switching process dynamics such as the expected duration, the probability of moving from one regime to another and the switching parameters of each regime. Finally, the model is able to accurately calibrate the parameters of drivers according to their driving regimes, so we can achieve a better understanding of drivers behavior and better simulation of driving situation. Index Terms car following model; regime classiﬁcation; model calibration; driver behavior; Markov switching model.

Computational docking is the core process of computer-aided drug design (CADD); it aims at predic... more Computational docking is the core process of computer-aided drug design (CADD); it aims at predicting the best orientation and conformation of a small molecule (drug ligand) when bound to a target large receptor molecule (protein) in order to form a stable complex molecule. The docking quality is typically measured by a scoring function: a mathematical predictive model that produces a score representing the binding free energy and hence the stability of the resulting complex molecule. An effective scoring function should produce promising drug candidates which can then be synthesized and physically screened using high throughput screening (HTS) process. Therefore, the key to CADD is the design of an efficient highly accurate scoring function. Many traditional techniques have been proposed, however, the performance was generally poor. Only in the last few years the application of the machine learning (ML) technology has been applied in the design of scoring functions; and the results have been very promising.

arXiv (Cornell University), Aug 23, 2016

EPJ Data Science

An in-Vehicle System and Method for During Accident Detection without being Fixed to Vehicle, 2020

System and Method for During Crash Accident Detection and Notification, 2020

2023 International Joint Conference on Neural Networks (IJCNN), 2023

Neural Computing and Applications

IEEE Geoscience and Remote Sensing Letters

ArXiv, 2021

Engineering Applications of Artificial Intelligence, 2015

Engineering Applications of Artificial Intelligence, 2014

Artificial Intelligence in Medicine, 2015

The objective of this article is to optimize the overall traffic flow on freeways using multiple ... more The objective of this article is to optimize the overall traffic flow on freeways using multiple ramp metering controls plus its complementary Dynamic Speed Limits (DSLs). An optimal freeway operation can be reached when minimizing the difference between the freeway density and the critical ratio for maximum traffic flow. In this article, a Multi-Agent Reinforcement Learning for Freeways Control (MARL-FWC) system for ramps metering and DSLs is proposed. MARL-FWC introduces a new microscopic framework at the network level based on collaborative Markov Decision Process modeling (Markov game) and an associated cooperative Q-learning algorithm. The technique incorporates payoff propagation (Max-Plus algorithm) under the coordination graphs framework, particularly suited for optimal control purposes. MARL-FWC provides three control designs: fully independent, fully distributed, and centralized; suited for different network architectures. MARL-FWC was extensively tested in order to assess the proposed model of the joint payoff, as well as the global payoff. Experiments are conducted with heavy traffic flow under the renowned VISSIM traffic simulator to evaluate MARL-FWC. The experimental results show a significant decrease in the total travel time and an increase in the average speed (when compared with the base case) while maintaining an optimal traffic flow. Keywords Multi-agent system · Intelligent transportation system · Adaptive traffic control · Freeway · Ramp metering · Max-plus · Q-learning · Sequential decision problem

Computational docking is the core process of computer-aided drug design; it aims at predicting th... more Computational docking is the core process of computer-aided drug design; it aims at predicting the best orientation and conformation of a small molecule (drug ligand) when bound to a target large receptor molecule (protein) in order to form a stable complex molecule. The docking quality is typically measured by a scoring function: a mathematical predictive model that produces a score representing the binding free energy and hence the stability of the resulting complex molecule. We analyze the performance of both learning techniques on the scoring power (binding affinity prediction), the ranking power (relative ranking prediction), docking power (identifying the native binding poses among computer-generated decoys), and screening power (classifying true binders versus negative binders) using the PDBbind 2013 database. For the scoring and ranking powers, the proposed learning scoring functions depend on a wide range of features (energy terms, pharmacophore, intermolecular) that entirely characterize the protein-ligand complexes (about 108 features); these features are extracted from several docking software available in the literature. For the docking and screening powers, the proposed learning scoring functions depend on the intermolecular features of the RF-Score (36 features) to utilize a larger number of training complexes (relative to the large number of decoys in the test set). For the scoring power, the DL RF scoring function (arithmetic mean between DL and RF scores) achieves Pearson's correlation coefficient between the predicted and experimentally measured binding affinities of 0.799 versus 0.758 of the RF scoring function. For the ranking power, the DL scoring function ranks the ligands bound to fixed target protein with accuracy 54% for the high-level ranking (correctly ranking the three ligands bound to the same target protein in a cluster) and with accuracy 78% for the low-level ranking (correctly ranking the best ligand only in the cluster) while the RF scoring function achieves (46% and 62%) respectively. For the docking power, the DL_RF scoring function has a success rate when the three best-scored ligand binding poses are considered within 2 ̊A root-mean-square-deviation from the native pose of 36.0% versus 30.2% of the RF scoring function. For the screening power, the DL scoring function has an average enrichment factor and success rate at the top 1% level of (2.69 and 6.45%) respectively versus (1.61 and 4.84%) respectively of the RF scoring function.

Machine learning (ML) is one of the fastest growing areas of science. It has been been used in ma... more Machine learning (ML) is one of the fastest growing areas of science. It has been been used in many applications; e.g., control problems , recommender systems, bioinformatics, etc. This is mainly attributed to the ability of machine learning techniques to utilize the current abundance in data; e.g., experimental, real-time, or on-line data. In addition, many machine learning techniques are able to solve problems where it is hard to formulate the problem in a closed form mathematical solution. As a result, some trials offer off-the-shelf packages for machine learning techniques, other trials have been offering domain-specific languages (DSL) for machine learning. However, the ability to add or extend such machine learning techniques is still limited. Thus, in this paper we highlight the ability of the minimization method to abstract the basic building blocks of such machine learning techniques. Minimization methods can solve the optimization problems numerically or symbolically. We highlight use cases on the ability of such minimization methods to solve some machine learning examples.

B.Sc. Thesis, 2005

In this thesis, we deal with a real world problem; that of buying and selling of electric power a... more In this thesis, we deal with a real world problem; that of buying and selling of electric power among distributed customers, resulting in a maximum reasonable proﬁt with the least possible effort. This is achieved by building a Multi Agent System (MAS), that implements Dutch auction format, such that each auction is composed of a number of auctioning and bidding agents along with a single transmission system operator agent that is used by the whole system. We implement the MAS using the open source library JADE (Java Agent DEvelopment Framework) through which we build agents that live on distributed platforms. The agents in the system communicate using FIPA-ACL (Foundation for Intelligent Physical Agents-Agent Communication Language) which is the most widely used and supported ACL. XML (eXtensible Markup Language) ﬁles are used to maintain customers’ accounts and various logging ﬁles.