Search

In this work, we study literature in Explainable AI and Safe AI to understand poisoning of neural models of code. In order to do so, we first establish a novel taxonomy for Trojan AI for code, and present a new aspect-based classification of triggers in neural models of code. Next, we highlight recent works that help us deepen our conception of how these models understand software code. Then we pick some of the recent, state-of-art poisoning strategies that can be used to manipulate such models. The insights we draw can potentially help to foster future research in the area of Trojan AI for code.

Although deep neural models substantially reduce the overhead of feature engineering, the features readily available in the inputs might significantly impact training cost and the performance of the models. In this paper, we explore the impact of an unsuperivsed feature enrichment approach based on variable roles on the performance of neural models of code. The notion of variable roles (as introduced in the works of Sajaniemi et al. [Refs. 1,2]) has been found to help students' abilities in programming. In this paper, we investigate if this notion would improve the performance of neural models of code. To the best of our knowledge, this is the first work to investigate how Sajaniemi et al.'s concept of variable roles can affect neural models of code. In particular, we enrich a source code dataset by adding the role of individual variables in the dataset programs, and thereby conduct a study on the impact of variable role enrichment in training the Code2Seq model. In addition, we shed light on some challenges and opportunities in feature enrichment for neural code intelligence models., Accepted in the 1st International Workshop on Interpretability and Robustness in Neural Software Engineering (InteNSE'23), Co-located with ICSE

Finding important features that contribute to the prediction of neural models is an active area of research in explainable AI. Neural models are opaque and finding such features sheds light on a better understanding of their predictions. In contrast, in this work, we present an inverse perspective of distractor features: features that cast doubt about the prediction by affecting the model's confidence in its prediction. Understanding distractors provide a complementary view of the features' relevance in the predictions of neural models. In this paper, we apply a reduction-based technique to find distractors and provide our preliminary results of their impacts and types. Our experiments across various tasks, models, and datasets of code reveal that the removal of tokens can have a significant impact on the confidence of models in their predictions and the categories of tokens can also play a vital role in the model's confidence. Our study aims to enhance the transparency of models by emphasizing those tokens that significantly influence the confidence of the models., The 1st International Workshop on Interpretability and Robustness in Neural Software Engineering, Co-located with ICSE (InteNSE'23)

The connectivity of public-safety mobile users (MU) in the co-existence of a public-safety network (PSN), unmanned aerial vehicles (UAVs), and LTE-based railway networks (LRN) needs a thorough investigation. UAVs are deployed as mobile base stations (BSs) for cell-edge coverage enhancement for MU. The co-existence of heterogeneous networks gives rise to the issue of co-channel interference due to the utilization of the same frequency band. By considering both sharing and non-sharing of radio access channels (RAC), we analyze co-channel interference in the downlink system of PSN, UAV, and LRN. As the LRN control signal demands high reliability and low latency, we provide higher priority to LRN users when allocating resources from the LRN RAC shared with MUs. Moreover, UAVs are deployed at the cell edge to increase the performance of cell-edge users. Therefore, interference control techniques enable LRN, PSN, and UAVs to cohabit in a scenario of sharing RAC. By offloading more PSN UEs to the LRN or UAVs, the resource utilization of the LRN and UAVs BSs is enhanced. In this paper, we aim to adopt deep learning (DL) based on enhanced inter-cell-interference coordination (eICIC) and further enhanced ICIC (FeICIC) strategies to deal with the interference from the PSN to the LRN and UAVs. Among LRN, PSN BS, and UAVs, a DL-based coordinated multipoint (CoMP) link technique is utilized to enhance the performance of PSN MUs. Therefore, if radio access channels are shared, utilization of DL-based FeICIC and CoMP for coordinated scheduling gives the best performance.

The growing prevalence of counterfeit stories on the internet has fostered significant interest towards fast and scalable detection of fake news in the machine learning community. While several machine learning techniques for this purpose have emerged, we observe that there is a need to evaluate the impact of noise on these techniques' performance, where noise constitutes news articles being mistakenly labeled as fake (or real). This work takes a step in that direction, where we examine the impact of noise on a state-of-the-art, structural model based on BiLSTM (Bidirectional Long-Short Term Model) for fake news detection, Hierarchical Discourse-level Structure for Fake News Detection by Karimi and Tang (Reference no. 9).

With the wide spread use of AI-driven systems in the edge (a.k.a edge intelligence systems), such as autonomous driving vehicles, wearable biotech devices, intelligent manufacturing, etc., such systems are becoming very critical for our day-to-day lives. A challenge in designing edge intelligence systems is that we have to deal with a large number of constraints in two design spaces that form the basis of such systems: the edge design space and the deep learning design space. Thus in this work, a new systematic, extendable, manual approach, READLE, is proposed for creating representations of specifications in edge intelligent systems, capturing constraints in the edge system design space (e.g. timing constraints and other performance constraints) and constraints in the deep learning space (e.g. model training duration, required level of accuracy) in a coherent fashion. In particular, READLE leverages benefits of real-time logic and binary decision diagrams to generate unified specifications. Several insights learned in building READLE are also discussed, which should help in future research in the domain of formal specifications for edge intelligent systems.

Software fuzzing mutates bytes in the test seeds to explore different behaviors of the program under test. Initial seeds can have great impact on the performance of a fuzzing campaign. Mutating a lot of uninteresting bytes in a large seed wastes the fuzzing resources. In this paper, we present the preliminary results of our approach that aims to improve the performance of fuzzers through identifying and removing uninteresting bytes in the seeds. In particular, we present DIAR, a technique that reduces the size of the seeds based on their coverage. Our preliminary results suggest fuzzing campaigns that start with reduced seeds, find new paths faster, and can produce higher coverage overall.

Software fuzzing is a strong testing technique that has become the de facto approach for automated software testing and software vulnerability detection in the industry. The random nature of fuzzing makes monitoring and understanding the behavior of fuzzers difficult. In this paper, we report the development of Fuzzer Mutation Visualizer (FMViz), a tool that focuses on visualizing byte-level mutations in fuzzers. In particular, FMViz extends American Fuzzy Lop (AFL) to visualize the generated test inputs and highlight changes between consecutively generated seeds as a fuzzing campaign progresses. The overarching goal of our tool is to help developers and students comprehend the inner-workings of the AFL fuzzer better. In this paper, we present the architecture of FMViz, discuss a sample case study of it, and outline the future work. FMViz is open-source and publicly available at https://github.com/AftabHussain/afl-test-viz.

The Facility for Rare Isotope Beams (FRIB) superconducting (SC) driver linac is designed to accelerate all stable ions including uranium to energies above 200 MeV/u primarily with 46 cryomodules containing 324 quarter-wave resonators (QWR) and half-wave (HWR) resonators. With the newly commissioned helium refrigeration system supplying liquid helium to the QWR and solenoids, heavy ion beams including Ne, Ar, Kr and Xe were accelerated to the charge stripper location above 20 MeV/u with the first linac segment consisting of 15 cryomodules containing 10⁴ QWRs of β=0.041 and 0.085 and 39 solenoids. Installation of cryomodules with β=0.29 and 0.53 HWRs is proceeding in parallel. Development of β=0.65 elliptical resonators is on-going supporting the FRIB energy upgrade to 400 MeV/u. This paper summarizes the SC-linac installation and phased commissioning status that is on schedule and on budget to the FRIB project., Proceedings of the 19th International Conference on RF Superconductivity, SRF2019, Dresden, Germany

The driver linac for Facility for Rare Isotope Beams (FRIB) will accelerate all stable ion beams from proton to uranium beyond 200 MeV/u with beam powers up to 400 kW. The linac now consists of 10⁴ superconducting quarter-wave resonators (QWR), which is the world largest number of low-beta SRF cavities operating at an accelerator facility. The first 3 QWR cryomodules (CM) (β = 0.041) were successfully integrated with cryogenics and other support systems for the 2nd Accelerator Readiness Review (ARR). The 3rd ARR scope that includes 11 QWR CM (β=0.085) and 1 QWR matching CM (β=0.085) was commissioned on schedule by January 2019, and then we met the Key Performance Parameters (KPP), accelerating Ar and Kr > 16 MeV/u at this stage, in a week upon the ARR authorization. We examine a variety of key factors to the successful commissioning, such as component testing prior to system integration, assessment steps of system/device readiness, and phased commissioning. This paper also reports on the integration process of the β=0.085 CMs including the test results, and the current progress on β=0.29 and 0.53 CMs in preparation for the upcoming 4th ARR., Proceedings of the North American Particle Accelerator Conference, NAPAC2019, Lansing, MI, USA

Facility for Rare Isotope Beams (FRIB) is a next-generation rare-isotope research facility under construction at Michigan State University (MSU). FRIB will produce rare-isotope beams of unprecedented intensities by impinging a 400 kW heavy-ion beam on a production target and by collecting and purifying the rare isotopes of interest with a fragment separator. A thermal imaging system (TIS) has been developed to monitor the beam spot on the production target. The main features and characteristics of optical system is presented. The prototype of optical system has been tested., Proceedings of the North American Particle Accelerator Conference, NAPAC2019, Lansing, MI, USA

In this work we present an account of the status of requirements engineering in the gaming industry. Recent papers in the area were surveyed. Characterizations of the gaming industry were deliberated upon by portraying its relations with the market industry. Some research directions in the area of requirements engineering in the gaming industry were also mentioned.

The Facility for Rare Isotope Beams (FRIB) driver linac is to accelerate all the stable ion beams from proton to uranium beyond 200 MeV/u with beam powers up to 400 kW, which will be the first large-scale, CW SRF ion linac. The beam commissioning of the front end (from the ion source to the RFQ) already began and is in progress. The Accelerator Readiness Review (ARR) for beam through the first three β=0.041 cryomodules is scheduled for May 2018. The next step is the beam commissioning through the 12 SRF cavities housed in these 3 cryomodules with 6 superconducting solenoid magnets. The cryomodules and the adjacent warm diagnostics boxes in between have been already installed and aligned in the tunnel. This paper describes the installation progress of the β=0.041 cryomodules and plans for ARR02., Proceedings of the 9th Int. Particle Accelerator Conf., IPAC2018, Vancouver, BC, Canada

With an average beam power two orders of magnitude higher than operating heavy-ion facilities, the Facility for Rare Isotope Beams (FRIB) stands at the power frontier of the accelerator family. This report summarizes the status of design, technology development, construction, commissioning, as well as path to operations and upgrades. We also highlight beam instrumentation challenges including machine protection of high-power heavy-ion beams and complications of multi-charge-state and multi-ion-species accelerations., Proceedings of the 6th Int. Beam Instrumentation Conf., IBIC2017, Grand Rapids, MI, USA

To evaluate the frequency and effects of blast-related otologic injuries.his retrospective study was conducted at the Aga Khan University Hospital, Karachi, and comprised charts of patients who were victims of bomb explosions between January 2011 and July 2013. Frequency and percentages were reported using cross tabulation with size of bomb, distance of person from blast and the presence of victim in open or closed space. Association of associated variables were also analysed.Of the 100 patients, 81(81%) were men and 19(19%) were women. Besides, 68(68%) patients were aged30 years. Also, 78(78%) subjects were exposed to80kg of explosives and 68(68%) were at a distance of10m. Furthermore, 61(61%) patients were exposed to explosion in openspace. The prevalence of ear injuries was 21(21%). The odds of experiencing various symptoms of ears was high in those who were exposed to80 kg of explosives (odds ratio: 3.38; 95% confidence interval, 1.16, 9.91). The odds of hearing loss in those who were within 10m was 8.62 (95% confidence interval: 2.72, 27.28) times than those who were10 m from the site of explosion.Otologic injuries were frequently associated with large blasts.

The FRIB Driver Linac* is to accelerate all the stable ions beyond 200 MeV/nucleon with a beam power of 400 kW. The linac is unique, being compactly folded twice. In this report, the room temperature magnets, amounting 147 in total, after Front End with a 0.5-MeV RFQ, are detailed, emphasizing the rotating coil field measurements and fiducialization., Proceedings of the North American Particle Accelerator Conf., NAPAC2016, Chicago, IL, USA

With an average beam power two orders of magnitude higher than operating heavy-ion facilities, the Facility for Rare Isotope Beams (FRIB) stands at the power frontier of the accelerator family. This report summarizes the current design and construction status as well as plans for commissioning, operations and upgrades., Proceedings of the 28th Linear Accelerator Conf., LINAC2016, East Lansing, MI, USA

With the advancement of silicon electronics under threat from physical limits to dimensional scaling, the International Technology Roadmap for Semiconductors (ITRS) released a white paper in 2008, detailing the ways in which the semiconductor industry can keep itself continually growing in the twenty-first century. Two distinct paths were proposed: More-Moore and More-than-Moore. While More-Moore approach focuses on the continued use of state-of-the-art, complementary metal oxide semiconductor (CMOS) technology for next generation electronics, More-than-Moore approach calls for a disruptive change in the system architecture and integration strategies. In this doctoral thesis, we investigate both the approaches to obtain performance improvement in the state-of-the-art, CMOS electronics. We present a novel channel material, SiSn, for fabrication of CMOS circuits. This investigation is in line with the More-Moore approach because we are relying on the established CMOS industry infrastructure to obtain an incremental change in the integrated circuit (IC) performance by replacing silicon channel with SiSn. We report a simple, low-cost and CMOS compatible process for obtaining single crystal SiSn wafers. Tin (Sn) is deposited on silicon wafers in the form of a metallic thin film and annealed to facilitate diffusion into the silicon lattice. This diffusion provides for sufficient SiSn layer at the top surface for fabrication of CMOS devices. We report a lowering of band gap and enhanced mobility for SiSn channel MOSFETs compared to silicon control devices. We also present a process for fabrication of vertically integrated flexible silicon to form 3D integrated circuits. This disruptive change in the state-of-the-art, in line with the More-than-Moore approach, promises to increase the performance per area of a silicon chip. We report a process for stacking and bonding these pieces with polymeric bonding and interconnecting them using copper through silicon vias (TSVs). We report a process for fabricating through polymer vias (TPVs) facilitating the fabrication of sensor arrays and control electronics on the opposite sides of the same flexible polymer. Finally, we present a process to fabricate stretchable metallic thin films with up to 800% stretchability, and report two distinct applications for these devices which cannot be done using current techniques.

The Facility for Rare Isotope Beams (FRIB) is based on a continuous-wave superconducting heavy ion linac to accelerate all the stable isotopes to above 200 MeV/u with a beam power of up to 400 kW. At an average beam power approximately two-to-three orders-of-magnitude higher than those of operating heavy-ion facilities, FRIB stands at the power frontier of the accelerator family - the first time for heavy-ion accelerators. In August 2014, the FRIB Project entered into full construction phase. Based on verified innovative designs, the FRIB accelerator team is working closely with partner laboratories and contracted industrial providers on the construction, installation and commissioning of the facility. This report summarizes the current design and construction status., Proceedings of the 13th Heavy Ion Accelerator Technology Conference, HIAT2015, Yokohama, Japan

This work focuses on the problem of high-performance robust control of a dc motor in the presence of a low-resolution position encoder. An observer assisted feedback control scheme is presented in which a robust extended high-gain observer (EHGO) is designed for simultaneous estimation of motor speed and lumped system uncertainties. The conventional approach of calculating speed by direct differentiation of encoder pulses is superseded by the EHGO technique that allows significant reduction in quantization noise associated with the encoder without requiring a low-pass filter. For the speed control problem, an improved integral sliding mode controller is proposed with a sigmoid boundary layer. The integral action compensates for the boundary layer error thereby guaranteeing tracking error convergence to zero. Stability, robustness and accuracy of the proposed observer and controller are established mathematically. Simulations and experimental tests are carried out that validate the given scheme.

Hussain et al. (2013) established new fixed point results in complete metric space. In this paper, we prove fixed point results of α-admissible mappings with respect to η, for modified contractive condition in complete metric space. An example is given to show the validity of our work. Our results generalize/improve several recent and classical results existing in the literature.

There has been an exponential increase in the performance of silicon based semiconductor devices in the past few decades. This improvement has mainly been due to dimensional scaling of the MOSFET. However, physical constraints limit the continued growth in device performance. To overcome this problem, novel channel materials are being developed to enhance carrier mobility and hence increase device performance. This work explores a novel semiconducting alloy - Silicon-tin (SiSn) as a channel material for CMOS applications. For the first time ever, MOS devices using SiSn as channel material have been demonstrated. A low cost, scalable and manufacturable process for obtaining SiSn by diffusion of Sn into silicon has also been explored. The channel material thus obtained is electrically characterized by fabricating MOSCAPs and Mesa-shaped MOSFETs. The SiSn devices have been compared to similar devices fabricated using silicon as channel material.

Geometries, electronic structure and electronic absorption spectra of thiophene based dye-sensitized solar cells were performed using Density Functional Theory (DFT) and time dependent density functional theory (TD-DFT). Different electron donating and electron withdrawing groups have been substituted. Geometries and electronic properties have been computed at B3LYP/6-31G ∗∗ and absorption spectra at TD-B3LYP/6-31G ∗∗ level of theory. Major change in bond lengths and bond angles occurs in the system where there is electron withdrawing or electron donating groups have been substituted. In SYSTEM-2 and SYSTEM-3 intra charge transfer has been observed. HOMO of SYSTEM-2 and SYSTEM-3 is delocalized on left side while LUMO on right side of the molecule. In SYSTEM-1, HOMO is on left side while LUMO is in the center. The designed systems show two absorption peaks for each of the system. In short, choice of appropriate electron withdrawing and donating groups is very important for improving the performance of dye-sensitized solar cells.