Collaborative activity recognition with heterogeneous activity sets and privacy preferences

One of the major challenges in Human Activity Recognition (HAR) based on machine learning is the scarcity of labeled data. Indeed, collecting a sufficient amount of training data to build a reliable recognition problem is often prohibitive. Among the many solutions in the literature to mitigate this issue, collaborative learning is emerging as a promising direction to distribute the annotation burden over multiple users that cooperate to build a shared recognition model. One of the major issues of existing methods is that they assume a static activity model with a fixed set of target activities. In this paper, we propose a novel approach that is based on Growing When Required (GWR) neural networks. A GWR network continuously adapts itself according to the input training data, and hence it is particularly suited when the users share heterogeneous sets of activities. Like in federated learning, for the sake of privacy preservation, each user contributes to the global activity classifier by sharing personal model parameters, and not by directly sharing data. In order to further mitigate privacy threats, we implement a strategy to avoid releasing model parameters that may indirectly reveal information about activities that the user specifically marked as private. Our results on two well-known publicly available datasets show the effectiveness and the flexibility of our approach.

Classification of Human Activity Recognition Utilizing Smartphone Data of CNN-LSTM

Human activity recognition has been applied in various areas of life by utilizing the gyroscope and accelerometer sensors embedded in smartphones. One of the functions of recognizing human activities is by understanding the pattern of human activity, thereby minimizing the possibility of unexpected incidents. This study classified of human activity recognition through CNN-LSTM on the UCI HAR dataset by applying the divide and conquer algorithm. This study additionally employs tuning hyperparameter to obtain the best accuracy value from the parameters and the proposed architecture. From the test results with the CNN-LSTM method, the accuracy rate for dynamic activity is 99.35%, for static activity is 96.08%, and the combination of the two models is 97.62%.

Discovery and Characterization of Low-Molecular Weight Inhibitors of Erwinia tracheiphila

Plant pathogenic bacteria in the genus Erwinia cause economically important diseases, including bacterial wilt of cucurbits caused by Erwinia tracheiphila. Conventional bactericides are insufficient to control this disease. Using high-throughput screening, 464 small molecules (SMs) with either cidal or static activity at 100 µM against a cucumber strain of E. tracheiphila were identified. Among them, 20 SMs (SM1 to SM20), composed of nine distinct chemical moiety structures, were cidal to multiple E. tracheiphila strains at 100 µM. These lead SMs had low toxicity to human cells and honey bees at 100 µM. No phytotoxicity was observed on melon plants at 100 µM, except when SM12 was either mixed with Silwet L-77 and foliar sprayed or when delivered through the roots. Lead SMs did not inhibit the growth of beneficial Pseudomonas and Enterobacter species but inhibited the growth of Bacillus species. Nineteen SMs were cidal to Xanthomonas cucurbitae and showed >50% growth inhibition against Pseudomonas syringae pv. lachrymans. In addition, 19 SMs were cidal or static against Erwinia amylovora in vitro. Five SMs demonstrated potential to suppress E. tracheiphila when foliar sprayed on melon plants at 2× the minimum bactericidal concentration. Thirteen SMs reduced Et load in melon plants when delivered via roots. Temperature and light did not affect the activity of SMs. In vitro cidal activity was observed after 3 to 10 h of exposure to these five SMs. Here, we report 19 SMs that provide chemical scaffolds for future development of bactericides against plant pathogenic bacterial species.

Non-Intrusive Presence Detection and Position Tracking for Multiple People Using Low-Resolution Thermal Sensors

This paper presents a framework to accurately and non-intrusively detect the number of people in an environment and track their positions. Different from most of the previous studies, our system setup uses only ambient thermal sensors with low-resolution, using no multimedia resources or wearable sensors. This preserves user privacy in the environment, and requires no active participation by the users, causing no discomfort. We first develop multiple methods to estimate the number of people in the environment. Our experiments demonstrate that algorithm selection is very important, but with careful selection, we can obtain up to 100% accuracy when detecting user presence. In addition, we prove that sensor placement plays a crucial role in the system performance, where placing the sensor on the room ceiling yields to the best results. After accurately finding the number of people in the environment, we perform position tracking on the collected ambient data, which are thermal images of the space where there are multiple people. We consider position tracking as static activity detection, where the user’s position does not change while performing activities, such as sitting, standing, etc. We perform efficient pre-processing on the data, including normalization and resizing, and then feed the data into well-known machine learning methods. We tested the efficiency of our framework (including the hardware and software setup) by detecting four static activities. Our results show that we can achieved up to 97.5% accuracy when detecting these static activities, with up to 100% class-wise precision and recall rates. Our framework can be very beneficial to several applications such as health-care, surveillance, and home automation, without causing any discomfort or privacy issues for the users.

The Effects of Innovative Shotgun Shooting Methods on Collegiate Shotgun Shooters

Sporting activities are classified according to movement demands and can be categorized as either dynamic or static actions. Many events exist within the discipline of “shooting sports”, and dynamic and static demands vary drastically among those events. However, consideration for differences in movement demands is frequently disregarded in shooting sports; common practice protocol encourages sh ooters to utilize static shooting techniques for all shooting sport events. In particular, shooting techniques for shotgun shooting, a dynamic sporting event, regularly align with rifle shooting (static activity) methods. Innovative dynamic shotgun shoot ing techniques have recently been developed, however, no previous studies have examined the outcomes of employing these dynamic techniques. Therefore, the current research investigated the effects of innovative shotgun shooting methods on collegiate shotg un shooters (n=38). Pre and post trap and skeet scores were collected at a certified International Shooting Sport Federation and USA Shooting competition field. Upon completion of pre-test shooting, subjects participated in an Optimum Shooting Performance (OSP) intervention that outlined innovative dynamic shooting and practice techniques. Post-test shooting scores were collected after2-weeks of OSP practice. A paired sample t test identified statistically significant improvements for trap shootingscores (t[32] = 2.82, p = .008, 95% CI [0.431, 2.660], d = .49),skeet shooting scores (t[32] = 2.59, p = .01, 95% CI[0.436, 3.625], d = .45), and total shooting (sum score of trap and skeet tests) scores (t[32] = 3.37, p = .002, 95%CI [1.417, 5.734], d = .59). These results suggest that learning and utilizing the OSP methods significantly increased the shooting performance of college shotgun shooters.

Spectrum-aware shared protection (SASP) algorithm for cognitive radio networks

Cognitive radio (CR) has become a key technology for addressing spectrum scarcity. In CR networks, spectrum access should not interfere the incumbent networks. Due to the requirement above, common control channel approaches, which are widely used in traditional multichannel environments, may face serious CR long-time blocking problem and control channel saturation problem. Although channel-hopping-based approaches can avoid these two problems, existing works still have significant drawbacks including long time-to-rendezvous, unbalance channel loading, and low channel utilization. This paper tends to the issue of range mindful survivable methodologies with disappointment likelihood limitations under static activity in adaptable transfer speed optical systems. The joint disappointment likelihood amongst essential and reinforcement ways must be beneath the most extreme fair joint disappointment likelihood for each activity request. It creates whole number direct program (ILP) models for committed way security and shared-way assurance with a specific end goal to limit the aggregate number of recurrence spaces expended, and furthermore propose a range mindful devoted insurance (SADP) calculation and a range mindful shared security (SASP) calculation. This demonstrates the ILP show arrangements devour least number of recurrence spaces, however prompt higher normal joint disappointment likelihood contrasted with the SADP and SASP calculations. In addition, both the SADP and SASP calculations accomplish a superior execution as far as aggregate number of recurrence openings expended when contrasted with a customary devoted way insurance calculation and an ordinary shared-way assurance calculation, separately, however prompt higher normal joint disappointment likelihood.

Maximum Memory Capacity on Neural Networks with Short-Term Synaptic Depression and Facilitation

In this work, we study, analytically and employing Monte Carlo simulations, the influence of the competition between several activity-dependent synaptic processes, such as short-term synaptic facilitation and depression, on the maximum memory storage capacity in a neural network. In contrast to the case of synaptic depression, which drastically reduces the capacity of the network to store and retrieve “static” activity patterns, synaptic facilitation enhances the storage capacity in different contexts. In particular, we found optimal values of the relevant synaptic parameters (such as the neurotransmitter release probability or the characteristic facilitation time constant) for which the storage capacity can be maximal and similar to the one obtained with static synapses, that is, without activity-dependent processes. We conclude that depressing synapses with a certain level of facilitation allow recovering the good retrieval properties of networks with static synapses while maintaining the nonlinear characteristics of dynamic synapses, convenient for information processing and coding.

The Namesake in Persian

Translation and literary translation in particular, no longer can be seen as a static activity. In Austermühl’s words, the antiquated image of a lone translator, armed only with a pencil or typewriter and surrounded by dusty books, is no longer realistic. The translators are now learning to apply the new communication technologies like Internet in their work. The web log as a recent phenomenon in the cyber world can be applied in literary translation. However, it has been overlooked in translation studies, partly, in author’s opinion, because of its novelty but to large extend, because of its popularity as a platform or a personal diary for the novice bloggers, the authors of the web logs. In this article, the application of a web log by an Iranian literary translator in his Persian translation of Jhumpa Lahiri’s The Namesake is examined. By providing examples of the discussions between the translator and his visitors (readers), and by incorporating the web log in a model of translation studies, The Holmes–Toury map of translation studies, he tries to pinpoint the advantages of web log in literary translation in respect to its application as an online translation workshop. The author argues that the web log can be thoroughly applied in literary translation provided that the translator optimizes the quality of his web log or workshop by acknowledging the growing importance of new technologies available for him; and making informed decisions for the effective use of translation technology. What distinguishes the online translation workshop from its traditional one is its round-the-clock opening, facilitated interaction with the translator, absence of any formal setting and formality and instant access to available sources in Internet. The application of weblogs as an online workshop has some advantages for the literary translator: acting as a mediator between the translator and his potential readers; sharing the experiences, problems and seeking advice; and, creating a virtual, universal world around the translation which is bigger than his small, personal world.