Search

Motional narrowing implies narrowing induced by motion, for example, in nuclear resonance, the thermally induced random motion of the nuclei in an inhomogeneous environment leads to counter-intuitive narrowing of the resonance line. Similarly, the excitons in monolayer semiconductors experience magnetic inhomogeneity: the electron-hole spin-exchange interaction manifests as an in-plane pseudo-magnetic field with a periodically varying orientation inside the exciton band. The excitons undergo random momentum scattering and pseudospin precession repeatedly in this inhomogeneous magnetic environment - typically resulting in fast exciton depolarization. On the contrary, we show that such magnetic inhomogeneity averages out at high scattering rate due to motional narrowing. Physically, a faster exciton scattering leads to a narrower pseudospin distribution on the Bloch sphere, implying a nontrivial improvement in exciton polarization. The in-plane nature of the pseudo-magnetic field enforces a contrasting scattering dependence between the circularly and linearly polarized excitons - providing a spectroscopic way to gauge the sample quality., Comment: Accepted in Nano Letters

In the recent past, significant research efforts have been put forth to fabricate low-cost noble metal-free substrates for surface-enhanced Raman spectroscopy (SERS) applications. Here we propose semiconducting TiO2 multi-leg nanotubes (TiO2 MLNTs, with and without the gold nanoparticle coating) as SERS substrates. TiO2 MLNTs show unique multi-leg morphology compared to the conventional non-multi-leg tubes and possess better light-harvesting properties. TiO2 MLNTs are fabricated with a simple and versatile single-step electrochemical anodization method. Remarkable high SERS sensitivity is observed towards the detection of Methylene blue (MB), up to nM concentration (E.F. ~104). The same is attributed to the resonantly matched photonic absorption edge of TiO2 MLNTs with the wavelength of incident laser probe light. On the other hand, gold nanoparticle-coated TiO2 MLNTs demonstrated further enhancement in SERS sensitivity (E.F. ~105, for nM of MB) facilitated by the synergy that exists between the plasmonic modes (LSPRs) of Au and the photonic absorption mode of TiO2 MLNTs.

The broadening and polarization of excitonic luminescence in monolayer TMDs largely suffer from inhomogeneity and temperature - an unresolved problem to date. In this work, through few-layer-graphene encapsulation of monolayer WS$_2$, we reduce the inter-excitonic energy separation, which then can have a narrow resonance with a specific phonon mode of our choice. The resulting single-step exciton relaxation with the resonating phonon mode significantly suppresses the inhomogeneous broadening, allowing us to achieve the narrowest exciton linewidth of 1.06 meV (which translates to 0.19 meV after deconvolution with the excitation laser linewidth). The single-phonon resonance helps to achieve a high quantum efficiency despite graphene encapsulation. The technique is powerful in tuning the exciton polarization during relaxation by choosing a specific resonating phonon mode. For example, the valley coherence (polarization) improves from $\sim$68% ($\sim$40%) to $\sim$90% ($\sim$75%) on resonance with 2$A_1$' and $A_1$' modes respectively. We further demonstrate a strong polarization reversal on resonance with a chiral phonon mode. Strikingly, the above features remain robust against temperature (up to 200 K) and sample age (few months in ambient condition). The findings will lead to clean excitonic measurements without requiring cryogenic cooling.

Strong Coulomb interaction in monolayer semiconductors allows them to host optically active large many-body states, such as the five-particle state, charged biexciton. Strong nonlinear light absorption by the charged biexciton under spectral resonance, coupled with its charged nature, makes it intriguing for nonlinear photodetection - an area that is hitherto unexplored. Using the high built-in vertical electric field in an asymmetrically designed few-layer graphene encapsulated 1L-WS$_2$ heterostructure, here we report a large, highly nonlinear photocurrent arising from the strong absorption by two charged biexciton species under zero external bias (self-powered mode). Time-resolved measurement reveals that the generated charged biexcitons transfer to the few-layer graphene in a timescale of sub-5 ps, indicating an ultrafast intrinsic limit of the photoresponse. By using single- and two-color photoluminescence excitation spectroscopy, we show that the two biexcitonic peaks originate from bright-dark and bright-bright exciton-trion combinations. Such innate nonlinearity in the photocurrent due to its biexcitonic origin, coupled with the ultrafast response due to swift inter-layer charge transfer, exemplifies the promise of manipulating many-body effects in monolayers towards viable optoelectronic applications.

Strong Coulomb interaction in monolayer transition metal dichalcogenides can facilitate nontrivial many-body effects among excitonic complexes. Many-body effects like exciton-exciton annihilation (EEA) have been widely explored in this material system. However, a similar effect for charged excitons (or trions), that is, trion-trion annihilation (TTA), is expected to be relatively suppressed due to repulsive like-charges, and has not been hitherto observed in such layered semiconductors. By a gate-dependent tuning of the spectral overlap between the trion and the charged biexciton through an "anti-crossing"-like behaviour in monolayer WS$_2$, here we present an experimental observation of an anomalous suppression of the trion emission intensity with an increase in gate voltage. The results strongly correlate with time-resolved measurements, and are inferred as a direct evidence of a nontrivial TTA resulting from non-radiative Auger recombination of a bright trion, and the corresponding energy resonantly promoting a dark trion to a charged biexciton state. The extracted Auger coefficient for the process is found to be tunable ten-fold through a gate-dependent tuning of the spectral overlap.

The monolayers of semiconducting transition metal dichalcogenides host strongly bound excitonic complexes and are an excellent platform for exploring many-body physics. Here we demonstrate a controlled kinetic manipulation of the five-particle excitonic complex, the charged biexciton, through a systematic dependence of the biexciton peak on excitation power, gate voltage, and temperature using steady-state and time-resolved photoluminescence (PL). With the help of a combination of the experimental data and a rate equation model, we argue that the binding energy of the charged biexciton is less than the spectral separation of its peak from the neutral exciton. We also note that while the momentum-direct radiative recombination of the neutral exciton is restricted within the light cone, such restriction is relaxed for a charged biexciton recombination due to the presence of near-parallel excited and final states in the momentum space.

Moir\'e superlattice-induced sub-bands in twisted van der Waals homo- and hetero-structures govern their optical and electrical properties, rendering additional degrees of freedom such as twist angle. Here, we demonstrate the moir\'e superlattice effects on the intralayer excitons and trions in a twisted bilayer of MoS$_2$ of H-type stacking at marginal twist angles. We identify the emission from localized and multiple delocalized sub-bands of intralayer moir\'e excitons and show their electrical modulation by the corresponding trion formation. The electrical control of the oscillator strength of the moir\'e excitons also results in a strong tunability of resonant Raman scattering. We find that the gate-induced doping significantly modulates the electronic moir\'e potential, however leaves the excitonic moir\'e confinement unaltered. This effect, coupled with variable moir\'e trap filling by tuning the optical excitation density, allows us to delineate the different phases of localized and delocalized moir\'e trions. We demonstrate that the moir\'e excitons exhibit strong valley coherence that changes in a striking non-monotonic W-shape with gating due to motional narrowing. These observations from the simultaneous electrostatic control of quasiparticle-dependent moir\'e potential will lead to exciting effects of tunable many-body phenomena in moir\'e superlattices.

Medical professionals evaluating alternative treatment plans for a patient often encounter time varying confounders, or covariates that affect both the future treatment assignment and the patient outcome. The recently proposed Counterfactual Recurrent Network (CRN) accounts for time varying confounders by using adversarial training to balance recurrent historical representations of patient data. However, this work assumes that all time varying covariates are confounding and thus attempts to balance the full state representation. Given that the actual subset of covariates that may in fact be confounding is in general unknown, recent work on counterfactual evaluation in the static, non-temporal setting has suggested that disentangling the covariate representation into separate factors, where each either influence treatment selection, patient outcome or both can help isolate selection bias and restrict balancing efforts to factors that influence outcome, allowing the remaining factors which predict treatment without needlessly being balanced., Comment: Accepted in Workshop on "The Neglected Assumptions in Causal Inference" at ICML 2021 (July)

In monolayer transition metal dichalcogenide semiconductors, valley coherence degrades rapidly due to a combination of fast scattering and inter-valley exchange interaction. This leads to a sub-picosecond valley coherence time, making coherent manipulation of exciton a highly challenging task. Using monolayer MoS2 sandwiched between top and bottom graphene, here we demonstrate fully valley coherent excitons by observing ~100% degree of linear polarization in steady state photoluminescence. This is achieved in this unique design through a combined effect of (a) suppression in exchange interaction due to enhanced dielectric screening, (b) reduction in exciton lifetime due to a fast inter-layer transfer to graphene, and (c) operating in the motional narrowing regime. We disentangle the role of the key parameters affecting valley coherence by using a combination of calculation (solutions of Bethe-Salpeter and Maialle-Silva-Sham equations) and a careful choice of design of experiments using four different stacks with systematic variation of screening and exciton lifetime. To the best of our knowledge, this is the first report in which the excitons are found to be valley coherent in the entire lifetime in monolayer semiconductors, allowing optical readout of valley coherence possible.

Advertising channels have evolved from conventional print media, billboards and radio advertising to online digital advertising (ad), where the users are exposed to a sequence of ad campaigns via social networks, display ads, search etc. While advertisers revisit the design of ad campaigns to concurrently serve the requirements emerging out of new ad channels, it is also critical for advertisers to estimate the contribution from touch-points (view, clicks, converts) on different channels, based on the sequence of customer actions. This process of contribution measurement is often referred to as multi-touch attribution (MTA). In this work, we propose CAMTA, a novel deep recurrent neural network architecture which is a casual attribution mechanism for user-personalised MTA in the context of observational data. CAMTA minimizes the selection bias in channel assignment across time-steps and touchpoints. Furthermore, it utilizes the users' pre-conversion actions in a principled way in order to predict pre-channel attribution. To quantitatively benchmark the proposed MTA model, we employ the real world Criteo dataset and demonstrate the superior performance of CAMTA with respect to prediction accuracy as compared to several baselines. In addition, we provide results for budget allocation and user-behaviour modelling on the predicted channel attribution., Comment: Accepted in ICDMW 2020

In 1H monolayer transition metal dichalcogenide, the inversion symmetry is broken, while the reflection symmetry is maintained. On the contrary, in the bilayer, the inversion symmetry is restored, but the reflection symmetry is broken. As a consequence of these contrasting symmetries, here we show that bilayer WS$_2$ exhibits a quantum confined Stark effect (QCSE) that is linear with the applied out-of-plane electric field, in contrary to a quadratic one for a monolayer. The interplay between the unique layer degree of freedom in the bilayer and the field driven partial inter-conversion between intra-layer and inter-layer excitons generates a giant tunability of the exciton oscillator strength. This makes bilayer WS$_2$ a promising candidate for an atomically thin, tunable electro-absorption modulator at the exciton resonance, particularly when stacked on top of a graphene layer that provides an ultra-fast non-radiative relaxation channel. By tweaking the biasing configuration, we further show that the excitonic response can be largely tuned through electrostatic doping, by efficiently transferring the oscillator strength from neutral to charged exciton. The findings are prospective towards highly tunable, atomically thin, compact and light, on chip, reconfigurable components for next generation optoelectronics., Comment: Accepted in ACS Photonics

We address the problem of counterfactual regression using causal inference (CI) in observational studies consisting of high dimensional covariates and high cardinality treatments. Confounding bias, which leads to inaccurate treatment effect estimation, is attributed to covariates that affect both treatments and outcome. The presence of high-dimensional co-variates exacerbates the impact of bias as it is harder to isolate and measure the impact of these confounders. In the presence of high-cardinality treatment variables, CI is rendered ill-posed due to the increase in the number of counterfactual outcomes to be predicted. We propose Hi-CI, a deep neural network (DNN) based framework for estimating causal effects in the presence of large number of covariates, and high-cardinal and continuous treatment variables. The proposed architecture comprises of a decorrelation network and an outcome prediction network. In the decorrelation network, we learn a data representation in lower dimensions as compared to the original covariates and addresses confounding bias alongside. Subsequently, in the outcome prediction network, we learn an embedding of high-cardinality and continuous treatments, jointly with the data representation. We demonstrate the efficacy of causal effect prediction of the proposed Hi-CI network using synthetic and real-world NEWS datasets., Comment: 23 pages, 5 figures, Accepted in Causal Discovery Workshop - KDD 2020

Causal inference (CI) in observational studies has received a lot of attention in healthcare, education, ad attribution, policy evaluation, etc. Confounding is a typical hazard, where the context affects both, the treatment assignment and response. In a multiple treatment scenario, we propose the neural network based MultiMBNN, where we overcome confounding by employing generalized propensity score based matching, and learning balanced representations. We benchmark the performance on synthetic and real-world datasets using PEHE, and mean absolute percentage error over ATE as metrics. MultiMBNN outperforms the state-of-the-art algorithms for CI such as TARNet and Perfect Match (PM)., Comment: 7 pages, 3 figures, Accepted in ESANN 2020

Performing inference on data obtained through observational studies is becoming extremely relevant due to the widespread availability of data in fields such as healthcare, education, retail, etc. Furthermore, this data is accrued from multiple homogeneous subgroups of a heterogeneous population, and hence, generalizing the inference mechanism over such data is essential. We propose the MetaCI framework with the goal of answering counterfactual questions in the context of causal inference (CI), where the factual observations are obtained from several homogeneous subgroups. While the CI network is designed to generalize from factual to counterfactual distribution in order to tackle covariate shift, MetaCI employs the meta-learning paradigm to tackle the shift in data distributions between training and test phase due to the presence of heterogeneity in the population, and due to drifts in the target distribution, also known as concept shift. We benchmark the performance of the MetaCI algorithm using the mean absolute percentage error over the average treatment effect as the metric, and demonstrate that meta initialization has significant gains compared to randomly initialized networks, and other methods., Comment: 10 pages, 4 figures, Accepted in CausalML Workshop - NeurIPS 2019

Layered transition metal dichalcogenides (TMDCs) host a variety of strongly bound exciton complexes that control the optical properties in these materials. Apart from spin and valley, layer index provides an additional degree of freedom in a few-layer thick film. Here we show that in a few-layer TMDC film, the wavefunctions of the conduction and valence band edge states contributing to the K (K') valley are spatially confined in the alternate layers - giving rise to direct (quasi-)intra-layer bright exciton and lower-energy inter-layer dark excitons. Depending on the spin and valley configuration, the bright exciton state is further found to be a coherent superposition of two layer-induced states, one (E-type) distributed in the even layers and the other (O-type) in the odd layers. The intra-layer nature of the bright exciton manifests as a relatively weak dependence of the exciton binding energy on the thickness of the few-layer film, and the binding energy is maintained up to 50 meV in the bulk limit - which is an order of magnitude higher than conventional semiconductors. Fast stokes energy transfer from the intra-layer bright state to the inter-layer dark states provides a clear signature in the layer-dependent broadening of the photoluminescence peak, and plays a key role in the suppression of the photoluminescence intensity observed in TMDCs with thickness beyond monolayer.

Monolayer Transition Metal Dichalcogenides (TMDS) are highly luminescent materials despite being sub-nanometer thick due to the ultra-short ($<1$ ps) radiative lifetime of the strongly bound bright excitons hosted by these materials. The intrinsically short radiative lifetime results in a large broadening in the exciton band with a magnitude that is about two orders greater than the spread of the light cone itself. The situation calls for a need to revisit the conventional light cone picture. We present a modified light cone concept which places the light line $(\hbar cQ)$ as the generalized lower bound for allowed radiative recombination. A self-consistent methodology, which becomes crucial upon inclusion of large radiative broadening in the exciton band, is proposed to segregate the radiative and the non-radiative components of the homogeneous exciton linewidth. We estimate a fundamental radiative linewidth of $1.54\pm0.17\ $meV, owing purely to finite radiative lifetime in the absence of non-radiative dephasing processes. As a direct consequence of the large radiative limit, we find a surprisingly large ($\sim 0.27 $ meV) linewidth broadening due to zero-point energy of acoustic phonons. This obscures the precise experimental determination of the intrinsic radiative linewidth and sets a fundamental limit on the non-radiative linewidth broadening at $T = 0$ K.

In this paper we present a comprehensive view of prominent causal discovery algorithms, categorized into two main categories (1) assuming acyclic and no latent variables, and (2) allowing both cycles and latent variables, along with experimental results comparing them from three perspectives: (a) structural accuracy, (b) standard predictive accuracy, and (c) accuracy of counterfactual inference. For (b) and (c) we train causal Bayesian networks with structures as predicted by each causal discovery technique to carry out counterfactual or standard predictive inference. We compare causal algorithms on two pub- licly available and one simulated datasets having different sample sizes: small, medium and large. Experiments show that structural accuracy of a technique does not necessarily correlate with higher accuracy of inferencing tasks. Fur- ther, surveyed structure learning algorithms do not perform well in terms of structural accuracy in case of datasets having large number of variables., Comment: arXiv admin note: text overlap with arXiv:1506.07669, arXiv:1611.03977 by other authors

Screening due to surrounding dielectric medium reshapes the electron-hole interaction potential and plays a pivotal role in deciding the binding energies of strongly bound exciton complexes in quantum confined monolayers of transition metal dichalcogenides (TMDs). However, owing to strong quasi-particle bandgap renormalization in such systems, a direct quantification of estimated shifts in binding energy in different dielectric media remains elusive using optical studies. In this work, by changing the dielectric environment, we show a conspicuous photoluminescence (PL) peak shift at low temperature for higher energy excitons (2s, 3s, 4s, 5s) in monolayer MoSe$_2$, while the 1s exciton peak position remains unaltered - a direct evidence of varying compensation between screening induced exciton binding energy modulation and quasi-particle bandgap renormalization. The estimated modulation of binding energy for the 1s exciton is found to be 58.6% (70.5% for 2s, 78.9% for 3s, 85% for 4s) by coating an Al$_2$O$_3$ layer on top, while the corresponding reduction in quasi-particle bandgap is estimated to be 248 meV. Such a direct evidence of large tunability of the binding energy of exciton complexes as well as the bandgap in monolayer TMDs holds promise of novel device applications., Comment: 19 pages including supplemental information

Monolayers of transition metal dichalcogenides (TMDCs) exhibit excellent electronic and optical properties. However, the performance of these two-dimensional (2D) devices are often limited by the large resistance offered by the metal contact interface. Till date, the carrier injection mechanism from metal to 2D TMDC layers remains unclear, with widely varying reports of Schottky barrier height (SBH) and contact resistance (Rc), particularly in the monolayer limit. In this work, we use a combination of theory and experiments in Au and Ni contacted monolayer MoS2 device to conclude the following points: (i) the carriers are injected at the source contact through a cascade of two potential barriers - the barrier heights being determined by the degree of interaction between the metal and the TMDC layer; (ii) the conventional Richardson equation becomes invalid due to the multi-dimensional nature of the injection barriers, and using Bardeen-Tersoff theory, we derive the appropriate form of the Richardson equation that describes such composite barrier; (iii) we propose a novel transfer length method (TLM) based SBH extraction methodology, to reliably extract SBH by eliminating any confounding effect of temperature dependent channel resistance variation; (iv) we derive the Landauer limit of the contact resistance achievable in such devices. A comparison of the limits with the experimentally achieved contact resistance reveals plenty of room for technological improvements., Comment: Accepted in Physical Review B

Aggregate analysis, such as comparing country-wise sales versus global market share across product categories, is often complicated by the unavailability of common join attributes, e.g., category, across diverse datasets from different geographies or retail chains, even after disparate data is technically ingested into a common data lake. Sometimes this is a missing data issue, while in other cases it may be inherent, e.g., the records in different geographical databases may actually describe different product 'SKUs', or follow different norms for categorization. Record linkage techniques can be used to automatically map products in different data sources to a common set of global attributes, thereby enabling federated aggregation joins to be performed. Traditional record-linkage techniques are typically unsupervised, relying textual similarity features across attributes to estimate matches. In this paper, we present an ensemble model combining minimal supervision using Bayesian network models together with unsupervised textual matching for automating such 'attribute fusion'. We present results of our approach on a large volume of real-life data from a market-research scenario and compare with a standard record matching algorithm. Finally we illustrate how attribute fusion using machine learning could be included as a data-lake management feature, especially as our approach also provides confidence values for matches, enabling human intervention, if required.

Discovering causal models from observational and interventional data is an important first step preceding what-if analysis or counterfactual reasoning. As has been shown before, the direction of pairwise causal relations can, under certain conditions, be inferred from observational data via standard gradient-boosted classifiers (GBC) using carefully engineered statistical features. In this paper we apply deep convolutional neural networks (CNNs) to this problem by plotting attribute pairs as 2-D scatter plots that are fed to the CNN as images. We evaluate our approach on the 'Cause- Effect Pairs' NIPS 2013 Data Challenge. We observe that a weighted ensemble of CNN with the earlier GBC approach yields significant improvement. Further, we observe that when less training data is available, our approach performs better than the GBC based approach suggesting that CNN models pre-trained to determine the direction of pairwise causal direction could have wider applicability in causal discovery and enabling what-if or counterfactual analysis., Comment: Published at NIPS 2016 Workshop "What If" and Won the best Poster Award

Today, cloud computing is an emerging way of computing in computer science. Cloud computing is a set of resources and services that are offered by the network or internet. Cloud computing extends various computing techniques like grid computing, distributed computing. Today cloud computing is used in both industrial field and academic field. Cloud facilitates its users by providing virtual resources via internet. As the field of cloud computing is spreading the new techniques are developing. This increase in cloud computing environment also increases security challenges for cloud developers. Users of cloud save their data in the cloud hence the lack of security in cloud can lose the users trust. In this paper we will discuss some of the cloud security issues in various aspects like multi-tenancy, elasticity, availability etc. The paper also discuss existing security techniques and approaches for a secure cloud. This paper will enable researchers and professionals to know about different security threats and models and tools proposed., Comment: 8 pages, 5 figures, SCNDS Ajmer