We present SUMO, a neural attention-based approach that learns to establish the correctness of textual claims based on evidence in the form of text documents (e.g., news articles or Web documents). SUMO further generates an extractive summary by pres
enting a diversified set of sentences from the documents that explain its decision on the correctness of the textual claim. Prior approaches to address the problem of fact checking and evidence extraction have relied on simple concatenation of claim and document word embeddings as an input to claim driven attention weight computation. This is done so as to extract salient words and sentences from the documents that help establish the correctness of the claim. However, this design of claim-driven attention does not capture the contextual information in documents properly. We improve on the prior art by using improved claim and title guided hierarchical attention to model effective contextual cues. We show the efficacy of our approach on datasets concerning political, healthcare, and environmental issues.
Measuring the congruence between two texts has several useful applications, such as detecting the prevalent deceptive and misleading news headlines on the web. Many works have proposed machine learning based solutions such as text similarity between
the headline and body text to detect the incongruence. Text similarity based methods fail to perform well due to different inherent challenges such as relative length mismatch between the news headline and its body content and non-overlapping vocabulary. On the other hand, more recent works that use headline guided attention to learn a headline derived contextual representation of the news body also result in convoluting overall representation due to the news bodys lengthiness. This paper proposes a method that uses inter-mutual attention-based semantic matching between the original and synthetically generated headlines, which utilizes the difference between all pairs of word embeddings of words involved. The paper also investigates two more variations of our method, which use concatenation and dot-products of word embeddings of the words of original and synthetic headlines. We observe that the proposed method outperforms prior arts significantly for two publicly available datasets.
Deep Neural Network (DNN) has gained unprecedented performance due to its automated feature extraction capability. This high order performance leads to significant incorporation of DNN models in different Internet of Things (IoT) applications in the
past decade. However, the colossal requirement of computation, energy, and storage of DNN models make their deployment prohibitive on resource constraint IoT devices. Therefore, several compression techniques were proposed in recent years for reducing the storage and computation requirements of the DNN model. These techniques on DNN compression have utilized a different perspective for compressing DNN with minimal accuracy compromise. It encourages us to make a comprehensive overview of the DNN compression techniques. In this paper, we present a comprehensive review of existing literature on compressing DNN model that reduces both storage and computation requirements. We divide the existing approaches into five broad categories, i.e., network pruning, sparse representation, bits precision, knowledge distillation, and miscellaneous, based upon the mechanism incorporated for compressing the DNN model. The paper also discussed the challenges associated with each category of DNN compression techniques. Finally, we provide a quick summary of existing work under each category with the future direction in DNN compression.
Development of future sensor, memory, and computing nanodevices based on novel physical concepts is one of the significant research endeavors in solid-state research. The field of spintronics is one such promising area of nanoelectronics which utiliz
es both the charge and spin of an electron for device operations. The advantage offered by spin systems is in their non-volatility and low-power functionality. This paper reviews emerging spintronic phenomena and the research advancements in diverse spin based applications. Spin devices and systems for logic, memories, emerging computing schemes, flexible electronics and terahertz emitters are discussed in this report.
Coronavirus outbreak is one of the most challenging pandemics for the entire human population of the planet Earth. Techniques such as the isolation of infected persons and maintaining social distancing are the only preventive measures against the epi
demic COVID-19. The actual estimation of the number of infected persons with limited data is an indeterminate problem faced by data scientists. There are a large number of techniques in the existing literature, including reproduction number, the case fatality rate, etc., for predicting the duration of an epidemic and infectious population. This paper presents a case study of different techniques for analysing, modeling, and representation of data associated with an epidemic such as COVID-19. We further propose an algorithm for estimating infection transmission states in a particular area. This work also presents an algorithm for estimating end-time of an epidemic from Susceptible Infectious and Recovered model. Finally, this paper presents empirical and data analysis to study the impact of transmission probability, rate of contact, infectious, and susceptible on the epidemic spread.
We present an {it ab initio}-based theoretical framework which elucidates the origin of the spin-orbit torque (SOT) in Normal-Metal(NM)/Ferromagnet(FM) heterostructures. The SOT is decomposed into two contributions, namely, {it spin-Hall} and the {it
spin-orbital} components. We find that {it (i)} the Field-Like (FL) SOT is dominated by the spin-orbital component and {it (ii)} both components contribute to the damping-like torque with comparable magnitude in the limit of thick Pt film. The contribution of the spin-orbital component to the DL-SOT is present only for NMs with strong SOC coupling strength. We demonstrate that the FL-SOT can be expressed in terms of the non-equilibrium spin-resolved orbital moment accumulation. The calculations reveal that the experimentally reported oxygen-induced sign-reversal of the FL-SOT in Pt/Co bilayers is due to the significant reduction of the majority-spin orbital moment accumulation on the interfacial NM atoms.
Ferromagnetic spintronics has been a main focus as it offers non-volatile memory and logic applications through current-induced spin-transfer torques. Enabling wider applications of such magnetic devices requires a lower switching current for a small
er cell while keeping the thermal stability of magnetic cells for non-volatility. As the cell size reduces, however, it becomes extremely difficult to meet this requirement with ferromagnets because spin-transfer torque for ferromagnets is a surface torque due to rapid spin dephasing, leading to the 1/ferromagnet-thickness dependence of the spin-torque efficiency. Requirement of a larger switching current for a thicker and thus more thermally stable ferromagnetic cell is the fundamental obstacle for high-density non-volatile applications with ferromagnets. Theories predicted that antiferromagnets have a long spin coherence length due to the staggered spin order on an atomic scale, thereby resolving the above fundamental limitation. Despite several spin-torque experiments on antiferromagnets and ferrimagnetic alloys, this prediction has remained unexplored. Here we report a long spin coherence length and associated bulk-like-torque characteristic in an antiferromagnetically coupled ferrimagnetic multilayer. We find that a transverse spin current can pass through > 10 nm-thick ferrimagnetic Co/Tb multilayers whereas it is entirely absorbed by 1 nm-thick ferromagnetic Co/Ni multilayer. We also find that the switching efficiency of Co/Tb multilayers partially reflects a bulk-like-torque characteristic as it increases with the ferrimagnet-thickness up to 8 nm and then decreases, in clear contrast to 1/thickness-dependence of Co/Ni multilayers. Our results on antiferromagnetically coupled systems will invigorate researches towards energy-efficient spintronic technologies.
Terahertz emission spectroscopy (TES) has recently played an important role in unveiling the spin dynamics at a terahertz (THz) frequency range. So far, ferromagnetic (FM)/nonmagnetic (NM) heterostructures have been intensively studied as THz sources
. Compensated magnets such as a ferrimagnet (FIM) and antiferromagnet (AFM) are other types of magnetic materials with interesting spin dynamics. In this work, we study TES from compensated magnetic heterostructures including CoGd FIM alloy or IrMn AFM layers. Systematic measurements on composition and temperature dependences of THz emission from CoGd/Pt bilayer structures are conducted. It is found that the emitted THz field is determined by the net spin polarization of the laser induced spin current rather than the net magnetization. The temperature robustness of the FIM based THz emitter is also demonstrated. On the other hand, an AFM plays a different role in THz emission. The IrMn/Pt bilayer shows negligible THz signals, whereas Co/IrMn induces sizable THz outputs, indicating that IrMn is not a good spin current generator, but a good detector. Our results not only suggest that a compensated magnet can be utilized for robust THz emission, but also provide a new approach to study the magnetization dynamics especially near the magnetization compensation point.
While current-induced spin-orbit torques (SOTs) have been extensively studied in ferromagnets and antiferromagnets, ferrimagnets have been less studied. Here we report the presence of enhanced spin-orbit torques resulting from negative exchange inter
action in ferrimagnets. The effective field and switching efficiency increase substantially as CoGd approaches its compensation point, giving rise to 9 times larger spin-orbit torques compared to that of non-compensated one. The macrospin modelling results also support efficient spin-orbit torques in a ferrimagnet. Our results suggest that ferrimagnets near compensation can be a new route for spin-orbit torque applications due to their high thermal stability and easy current-induced switching assisted by negative exchange interaction.
We theoretically study the influence of a predominant field-like spin-orbit torque on the magnetization switching of small devices with a uniform magnetization. We show that for a certain range of ratios (0.23-0.55) of the Slonczewski to the field-li
ke torques, it is possible to deterministically switch the magnetization without requiring any external assist field. A precise control of the pulse length is not necessary, but the pulse edge sharpness is critical. The proposed switching scheme is numerically verified to be effective in devices by micromagnetic simulations. Switching without any external assist field is of great interest for the application of spin-orbit torques to magnetic memories.
