Reservoir Computing (RC) offers a viable option to deploy AI algorithms on low-end embedded system platforms. Liquid State Machine (LSM) is a bio-inspired RC model that mimics the cortical microcircuits and uses spiking neural networks (SNN) that can
be directly realized on neuromorphic hardware. In this paper, we present a novel Parallelized LSM (PLSM) architecture that incorporates spatio-temporal read-out layer and semantic constraints on model output. To the best of our knowledge, such a formulation has been done for the first time in literature, and it offers a computationally lighter alternative to traditional deep-learning models. Additionally, we also present a comprehensive algorithm for the implementation of parallelizable SNNs and LSMs that are GPU-compatible. We implement the PLSM model to classify unintentional/accidental video clips, using the Oops dataset. From the experimental results on detecting unintentional action in video, it can be observed that our proposed model outperforms a self-supervised model and a fully supervised traditional deep learning model. All the implemented codes can be found at our repository https://github.com/anonymoussentience2020/Parallelized_LSM_for_Unintentional_Action_Recognition.
The Higgs boson may well be a composite scalar with a finite extension in space. Owing to the momentum dependence of its couplings the imprints of such a composite pseudo Goldstone Higgs may show up in the tails of various kinematic distributions at
the LHC, distinguishing it from an elementary state. From the bottom up we construct the momentum dependent form factors to capture the interactions of the composite Higgs boson with the weak gauge bosons. We demonstrate their impact in the differential distributions of various kinematic parameters for the $pprightarrow Z^*Hrightarrow l^+l^-bbar{b}$ channel. We show that this channel can provide an important avenue to probe the Higgs substructure at the HL-LHC.
Analysis of human sketches in deep learning has advanced immensely through the use of waypoint-sequences rather than raster-graphic representations. We further aim to model sketches as a sequence of low-dimensional parametric curves. To this end, we
propose an inverse graphics framework capable of approximating a raster or waypoint based stroke encoded as a point-cloud with a variable-degree Bezier curve. Building on this module, we present Cloud2Curve, a generative model for scalable high-resolution vector sketches that can be trained end-to-end using point-cloud data alone. As a consequence, our model is also capable of deterministic vectorization which can map novel raster or waypoint based sketches to their corresponding high-resolution scalable Bezier equivalent. We evaluate the generation and vectorization capabilities of our model on Quick, Draw! and K-MNIST datasets.
This work presents a Cross-device Deep-Learning based Electromagnetic (EM-X-DL) side-channel analysis (SCA), achieving >90% single-trace attack accuracy on AES-128, even in the presence of significantly lower signal-to-noise ratio (SNR), compared to
the previous works. With an intelligent selection of multiple training devices and proper choice of hyperparameters, the proposed 256-class deep neural network (DNN) can be trained efficiently utilizing pre-processing techniques like PCA, LDA, and FFT on the target encryption engine running on an 8-bit Atmel microcontroller. Finally, an efficient end-to-end SCA leakage detection and attack framework using EM-X-DL demonstrates high confidence of an attacker with <20 averaged EM traces.
It has long been recognized that academic success is a result of both cognitive and non-cognitive dimensions acting together. Consequently, any intelligent learning platform designed to improve learning outcomes (LOs) must provide actionable inputs t
o the learner in these dimensions. However, operationalizing such inputs in a production setting that is scalable is not trivial. We develop an Embibe Score Quotient model (ESQ) to predict test scores based on observed academic, behavioral and test-taking features of a student. ESQ can be used to predict the future scoring potential of a student as well as offer personalized learning nudges, both critical to improving LOs. Multiple machine learning models are evaluated for the prediction task. In order to provide meaningful feedback to the learner, individualized Shapley feature attributions for each feature are computed. Prediction intervals are obtained by applying non-parametric quantile regression, in an attempt to quantify the uncertainty in the predictions. We apply the above modelling strategy on a dataset consisting of more than a hundred million learner interactions on the Embibe learning platform. We observe that the Median Absolute Error between the observed and predicted scores is 4.58% across several user segments, and the correlation between predicted and observed responses is 0.93. Game-like what-if scenarios are played out to see the changes in LOs, on counterfactual examples. We briefly discuss how a rational agent can then apply an optimal policy to affect the learning outcomes by treating the above model like an Oracle.
Surface bound catalytic chemical reactions self-propel chemically active Janus particles. In the vicinity of boundaries, these particles exhibit rich behavior, such as the occurrence of wall-bound steady states of sliding. Most active particles tend
to sediment as they are density mismatched with the solution. Moreover Janus spheres, which consist of an inert core material decorated with a cap-like, thin layer of a catalyst, are gyrotactic (bottom-heavy). Occurrence of sliding states near the horizontal walls depends on the interplay between the active motion and the gravity-driven sedimentation and alignment. It is thus important to understand and quantify the influence of these gravity-induced effects on the behavior of model chemically active particles moving near walls. For model gyrotactic, self-phoretic Janus particles, here we study theoretically the occurrence of sliding states at horizontal planar walls that are either below (floor) or above (ceiling) the particle. We construct state diagrams characterizing the occurrence of such states as a function of the sedimentation velocity and of the gyrotactic response of the particle, as well as of the phoretic mobility of the particle. We show that in certain cases sliding states may emerge simultaneously at both the ceiling and the floor, while the larger part of the experimentally relevant parameter space corresponds to particles that would exhibit sliding states only either at the floor or at the ceiling or there are no sliding states at all. These predictions are critically compared with the results of previous experimental studies and our experiments conducted on Pt-coated polystyrene and silica-core particles suspended in aqueous hydrogen peroxide solutions.
In this work we reappraise the collider constraints from leptonic final states on the vectorlike colored top partners taking into account the impact of exotic colored vector resonances. These colored states are intrinsic to a broad class of models th
at employ a strongly interacting sector to drive electroweak symmetry breaking. We translate the recent results in the {sl monolepton + jets} channel as reported by CMS with an integrated luminosity of 35.8 fb$^{-1}$, and {sl dilepton + jets} and {sl trilepton + jets} channels as reported by ATLAS with an integrated luminosity of 36.1 fb$^{-1}$ to constrain the parameter space of these class of models. We also comment on the impact and modification of the derived constraints due to the expected fatness of the colored vector resonance, when accounted for beyond the narrow-width approximation by simulating the full one-particle irreducible resummed propagator.
Biomedical researchers usually study the effects of certain exposures on disease risks among a well-defined population. To achieve this goal, the gold standard is to design a trial with an appropriate sample from that population. Due to the high cost
of such trials, usually the sample size collected is limited and is not enough to accurately estimate some exposures effect. In this paper, we discuss how to leverage the information from external `big data (data with much larger sample size) to improve the estimation accuracy at the risk of introducing small bias. We proposed a family of weighted estimators to balance the bias increase and variance reduction when including the big data. We connect our proposed estimator to the established penalized regression estimators. We derive the optimal weights using both second order and higher order asymptotic expansions. Using extensive simulation studies, we showed that the improvement in terms of mean square error (MSE) for the regression coefficient can be substantial even with finite sample sizes and our weighted method outperformed the existing methods such as penalized regression and James Steins approach. Also we provide theoretical guarantee that the proposed estimators will never lead to asymptotic MSE larger than the maximum likelihood estimator using small data only in general. We applied our proposed methods to the Asia Cohort Consortium China cohort data to estimate the relationships between age, BMI, smoking, alcohol use and mortality.
Optimal biomarker combinations for treatment-selection can be derived by minimizing total burden to the population caused by the targeted disease and its treatment. However, when multiple biomarkers are present, including all in the model can be expe
nsive and hurt model performance. To remedy this, we consider feature selection in optimization by minimizing an extended total burden that additionally incorporates biomarker measurement costs. Formulating it as a 0-norm penalized weighted classification, we develop various procedures for estimating linear and nonlinear combinations. Through simulations and a real data example, we demonstrate the importance of incorporating feature-selection and marker cost when deriving treatment-selection rules.
In this paper, we study two issues in asynchronous communication systems. The first issue is the derivation of sum capacity bounds for finite dimensional asynchronous systems. In addition, asymptotic results for the sum capacity bounds are obtained.
The second issue is the design of practical suboptimal codes for binary chip asynchronous CDMA systems that become optimal for high Signal-to-Noise (SNR) ratios. The performance of such suboptimal codes is also compared to Gold and Optical Orthogonal codes. The conclusion is that the proposed suboptimal codes perform favorably compared to other known codes for high SNR asynchronous systems and perform more or less the same as the other codes for the low SNR values.
