The convolutional neural network (CNN)-based models have achieved tremendous breakthroughs in many end-to-end applications, such as image identification, text classification, and speech recognition. By replicating these successes to the field of malware detection, several CNN-based malware detectors have achieved encouraging performance without significant feature engineering effort in recent years. Unfortunately, by analyzing their robustness using gradient-based algorithms, several studies have shown that some of these malware detectors are vulnerable to the evasion attacks (also known as adversarial examples). However, the existing attack methods can only achieve quite low attack success rates. In this paper, we propose two novel white-box methods and one novel black-box method to attack a recently proposed malware detector. By incorporating the gradient-based algorithm, one of our white-box methods can achieve a success rate of over 99%. Without prior knowledge of the exact structure and internal parameters of the detector, the proposed black-box method can also achieve a success rate of over 70%. In addition, we consider adversarial training as a defensive mechanism in order to resist evasion attacks. While proving the effectiveness of adversarial training, we also analyze its security risk, that is, a large number of adversarial examples can poison the training dataset of the detector. Therefore, we propose a pre-detection mechanism to reject adversarial examples. The experiments show that this mechanism can effectively improve the safety and efficiency of malware detection.

Cutting forces in machining process provide useful information in understanding the mechanics of machining process, tool wear, tool/workpiece material selection, and quality of a machined surface. In addition, cutting force measurements has become a crucial activity for process enhancement and optimization. In this study, a strain gauges-based novel force dynamometer capable of measuring cutting forces in facing process has been designed and manufactured to measure optimal cutting parameters for the studied material, i.e., mild steel A1010. The selection of orientation of strain gauges was set in the developed dynamometer to have maximum sensitivity and minimum cross-sensitivity during facing process. The dynamometer was connected to a quarter bridge data acquisition system for signal capturing and processing to achieve cutting forces at selected cutting conditions. The rigidity and stiffness of the dynamometer were also analyzed by determining its natural frequency at the design stage. Finally, Taguchi method is deployed on experimental results at specified cutting parameters to get optimal parameters for selected material. The dynamometer was experimentally tested, and results obtained are found in good relation to the numerical data (simulated) that confirm its reliability to measure the cutting forces in all three components (x, y, and z) for facing process.

The Internet of Things (IoT) has grown rapidly in recent years and has become one of the most active areas in the global market. As an open source platform with a large number of users, Android has become the driving force behind the rapid development of the IoT, also attracted malware attacks. Considering the explosive growth of Android malware in recent years, there is an urgent need to propose efficient methods for Android malware detection. Although the existing Android malware detection methods based on machine learning has achieved encouraging performance, most of these methods require a lot of time and effort from the malware analysts to build dynamic or static features, so these methods are difficult to apply in practice. Therefore, end-to-end malware detection methods without human expert intervention are required. This paper proposes two end-to-end Android malware detection methods based on deep learning. Compared with the existing detection methods, the proposed methods have the advantage of their end-to-end learning process. Our proposed methods resample the raw bytecodes of the classes.dex files of Android applications as input to deep learning models. These models are trained and evaluated in a dataset containing 8K benign applications and 8K malicious applications. Experiments show that the proposed methods can achieve 93.4% and 95.8% detection accuracy respectively. Compared with the existing methods, our proposed methods are not limited by input filesize, no manual feature engineering, low resource consumption, so they are more suitable for application on Android IoT devices. (C) 2020 Elsevier B.V. All rights reserved.

Snow leopard (Panthera uncia) is an elusive endangered carnivore found in remote mountain regions of Central Asia, with sparse distribution in northern Pakistan, including Chitral and Baltistan. The present study determined the food habits of snow leopard, including preferred prey species and seasonal variation in diet. Fifty-six scat samples were collected and analyzed to determine the diet composition in two different seasons, i.e. summer and winter. Hair characteristics such as cuticular scale patterns and medullary structure were used to identify the prey. This evidence was further substantiated from the remains of bones, claws, feathers, and other undigested remains found in the scats. A total of 17 prey species were identified; 5 of them were large mammals, 6 were mesomammals, and the remaining 6 were small mammals. The occurrence of wild ungulates (10.4%) in the diet was low, while livestock constituted a substantial part (26.4%) of the diet, which was higher in summer and lower in winter. Mesomammals altogether comprised 33.4% of the diet, with palm civet (Paguma larvata) as a dominant (16.8%) species, followed by golden marmot (Marmota caudate) (8.8%), which was higher in winter. There was a significant difference in seasonal variation in domestic livestock and small mammals. The livestock contribution of 26.4% observed in the present study indicates a significant dependence of the population on livestock and suggests that the study area is expected to be a high-conflict area for snow leopards. The results of the current study would help improve the conservation efforts for snow leopards, contributing to conflict resolution and effective management of this endangered cat.

Optimization of a manufacturing process results in higher productivity and reduced wastes. Production parameters of a local steel bar manufacturing industry of Pakistan is optimized by using six Sigma-Define, measure, analyze, improve, and control methodology. Production data is collected and analyzed. After analysis, experimental design result is used to identify significant factors affecting process performance. The significant factors are controlled to optimized level using two-level factorial design method. A regression model is developed that helps in the estimation of response under multi variable input values. Model is tested, verified, and validated by using industrial data collected at a local steel bar manufacturing industry of Peshawar(Khyber Pakhtunkhwa, Pakistan). The sigma level of the manufacturing process is improved to 4.01 from 3.58. The novelty of the research is the identification of the significant factors along with the optimum levels that affects the process yield, and the methodology to optimize the steel bar manufacturing process.

Carpooling enables commuters to share travel expenses, save costs, and improve their mobility options and reduces emission and traffic congestion. To commute by carpooling, individuals need to communicate, negotiate, and coordinate, and in most cases they need to adapt their schedule to enable cooperation. This paper presents the design of an agent-based model by defining phases and steps that may be taken to move from solo driving to carpooling. The paper analyzes the various effects of agent interaction and behavior adaptation for a set of candidate carpoolers. The start of the carpooling process depends on the individuals' objectives and intention to carpool. Through negotiation and coordination, individuals can reach complex agreements in an iterative way. The success of negotiation highly depends on the lifestyle factors that influence the departure time decision, on the profile of the individuals, and on the effect of constraining activities. The carpooling social network was established by use of the results predicted by FEATHERS, an operational activity based model for Flanders, Belgium. From the simulation's discussions, it is possible to portray the true picture of potential carpoolers throughout their carpooling period. The simulation results show that 9.33% of the commuters started to carpool when the time window was +/- 30 min and the average occupancy per car was 2.4 persons. When the time window was larger, the chances for negotiation success were greater than those when a smaller time window was used. Hence, carpooling requires time flexibility. The Janus (multiagent) platform was used to simulate the interactions of autonomous agents.

Modeling the interaction between individual agents becomes progressively important in recent research. Carpooling for commuters is a specific transportation problem where cooperation between agents is essential while executing their daily schedule. Organization-based modeling provides the ability to determine where the relationships between agents exist and how these relationships influence the results. This paper presents both the design of an organizational model that is mapped to an agent based simulation model and a proof of concept implementation. It analyzes various effects of agent interaction and behavior adaptation for sets of candidate carpoolers. The goal is to limit the interactions of autonomous agents, to enable communication to trigger the negotiation process within social groups. The start of the carpooling process depends on the individuals' objectives and intention to carpool. The success of negotiation highly depends on the trip departure time preference, on the individuals' profile, route optimization and on the effect of constraining activities. In order to cooperate individuals adapt their agenda according to personal preferences and limitations. The carpooling social network was established using results predicted by the FEATHERS operational activity-based model for Flanders (Belgium). From the simulation's discussions, it is possible to portray the real picture of the potential carpoolers throughout their carpooling period. The Janus (multi-agent) platform is used for simulating the interactions of autonomous individuals. (C) 2016 Elsevier B.V. All rights reserved.

In the carpooling, individuals need to communicate, negotiate and in most cases adapt their daily schedule to enable cooperation. Through negotiation, agents (individuals) can reach complex agreements in an iterative way which meet the criteria for successful negotiation. The result of the negotiation depends on "negotiation mechanism" used to match and on the behavior of the agents involved in the negotiation process. This paper presents an organizational and agent-based model for commuting by candidate carpoolers using a simple negotiation mechanism aimed at finding an acceptable agreement between agents to carpool. Initially, the agents involved in exploration process, search for their partners via some kind of Agent Communication Language (ACL); after finding potential partners, they start a negotiation to find matched partner to carpool. After having found a good match, the agents can carpool for a specified time period. The agents join the carpool group when the negotiation is successful and leave the carpool group when the agreed time period is expired. Agents can be part of several carpool groups sequentially. The first implementation used home and work locations as well as preferred trip start times and carpool periods determined by uniformly sampling given sets. Furthermore a simplistic negotiation mechanism used roughly to produce possible results for the synthetic data. An automated negotiation model is implemented and validated through simulation. The Janus multi-agent platform is used. Future research will mainly focus on the development of behaviorally sound negotiation mechanism. (C) 2014 The Authors. Published by Elsevier B.V.

The Indian pangolin (Manis crassicaudata) has been recently listed by the International Union for Conservation of Nature as an endangered species throughout its range, but in Pakistan it is categorized as vulnerable. Very little is known about the breeding habits of this nocturnal and fossorial animal in the wild. The present paper provides information on breeding ecology of its population in Potohar Plateau in Pakistan. A total of 13 specimens were trapped to record breeding condition of the captured animals. Additionally, a questionnaire survey was conducted in the study area to collect breeding data on the species. Our limited data show a male-to-female ratio of 1.6:1. The local population seems to breed once a year, usually from July to October, with a litter size of one to two. The juvenile pangolins were observed during the months of January, April and December.

Knowledge management success is a hot issue in SMEs. It is obvious that several important factors must be considered for successful implementation, but most small and medium firms have no idea what factors should be considered most heavily. Therefore, the purpose of this study is to expand the base of knowledge in that area, and empirically test the relationship between personal capabilities within SMEs and knowledge management system success. Six factors related to personal capabilities were measured: ambition, skills, behaviour, tools and techniques, time management and personal knowledge. The results of the study indicate that there is a significant relationship between these factors and knowledge management system success.