arXiv subjects

Tarun Agarwal

Publications and source records attributed to Tarun Agarwal.

5 recordsLinked to original sources

arXiv2025

A Representation Sharpening Framework for Zero Shot Dense Retrieval

Zero-shot dense retrieval is a challenging setting where a document corpus is provided without relevant queries, necessitating a reliance on pretrained dense retrievers (DRs). However, since these DRs are not trained on the target corpus, they struggle to represent semantic differences between similar documents. To address this failing, we introduce a training-free representation sharpening framework that augments a document's representation with information that helps differentiate it from similar documents in the corpus. On over twenty datasets spanning multiple languages, the representation sharpening framework proves consistently superior to traditional retrieval, setting a new state-of-the-art on the BRIGHT benchmark. We show that representation sharpening is compatible with prior approaches to zero-shot dense retrieval and consistently improves their performance. Finally, we address the performance-cost tradeoff presented by our framework and devise an indexing-time approximation that preserves the majority of our performance gains over traditional retrieval, yet suffers no additional inference-time cost.

cs.IR↗

arXiv2022

Dynamics of van der Waals Charge Qubit in 2D Bilayers: Ab initio Quantum Transport and Qubit Measurement

A van der Waals (vdW) charge qubit, electrostatically confined within two-dimensional (2D) vdW materials, is proposed as building block of future quantum computers. Its characteristics are systematically evaluated with respect to its two-level anti-crossing energy difference ($Δ$). Bilayer graphene ($Δ$ $\approx$ 0) and a vdW heterostructure ($Δ$ $\gg$ 0) are used as representative examples. Their tunable electronic properties with an external electric field define the state of the charge qubit. By combining density functional theory and quantum transport calculations, we highlight the optimal qubit operation conditions based on charge stability and energy-level diagrams. Moreover, a single-electron transistor (SET) design based on trilayer vdW heterostructures capacitively coupled to the charge qubit is introduced as measurement setup with low decoherence and improved measurement properties. It is found that a $Δ$ greater than 20 meV results in a rapid mixing of the qubit states, which leads to a lower measurement quantity, i.e. contrast and conductance. With properly optimized designs, qubit architectures relying on 2D vdW structures could be integrated into an all-electronic quantum computing platform.

cond-mat.mes-hall↗

arXiv2021

Circuit-aware Device Modeling of Energy-efficient Monolayer WS$_2$ Trench-FinFETs

The continuous scaling of semiconductor technology has pushed the footprint of logic devices below 50 nm. Currently, logic standard cells with one single fin are being investigated to increase the integration density, although such options could severely limit the performance of individual devices. In this letter, we present a novel Trench (T-) FinFET device, composed of a monolayer two-dimensional (2D) channel material. The device characteristics of a monolayer WS$_2$-based T-FinFET are studied by combining the first-principles calculations and quantum transport (QT) simulations. These results serve as inputs to a predictive analytical model. The latter allows to benchmark the T-FinFET with strained (s)-Si FinFETs in both quasi-ballistic and diffusive transport regimes. The circuit-level evaluation highlights that WS$_2$ T-FinFETs exhibit a competitive energy-delay performance compared to s-Si FinFET and WS$_2$ double-gate transistors, assuming the same mobility and contact resistivity at small footprints.

cond-mat.mtrl-sci↗

arXiv2019

On-Chip Concentration and Patterning of Biological Cells Using Interplay of Electrical and Thermal Fields

We demonstrate a method of concentrating and patterning of biological cells on a chip, exploiting the confluence of electric and thermal fields, without necessitating the use of any external heating or illuminating source. The technique simply employs two parallel plate electrodes and an insulating layer over the bottom electrode, with a drilled insulating layer for inducing localized variations in the thermal field. A strong induced electric field, in the process, penetrates through the narrow hole and generates highly non-uniform heating, which in turn, results in gradients in electrical properties and induces mobile charges to impose directional fluid flow. The toroidal vortices, induced by secondary electrokinetic forces originating out of temperature-dependent electrical property variations, transport the suspended cells towards a hot-spot site of the chip, for rapid concentrating and patterning into different shaped clusters based on pre-designed conditions, without exceeding safe temperature limits that do not result in damage of thermally labile biological samples. We characterize the efficacy of the cell trapping process for two different biological entities, namely, Escherichia coli bacteria and yeast cell. These results may be of profound importance towards developing novel biomedical microdevices for drug discovery, antibiotic resistance assessment and medical diagnostics.

physics.bio-ph↗

arXiv2011

Noncooperative Games for Autonomous Consumer Load Balancing over Smart Grid

Traditionally, most consumers of electricity pay for their consumptions according to a fixed rate. With the advancement of Smart Grid technologies, large-scale implementation of variable-rate metering becomes more practical. As a result, consumers will be able to control their electricity consumption in an automated fashion, where one possible scheme is to have each individual maximize its own utility as a noncooperative game. In this paper, noncooperative games are formulated among the electricity consumers in Smart Grid with two real-time pricing schemes, where the Nash equilibrium operation points are investigated for their uniqueness and load balancing properties. The first pricing scheme charges a price according to the average cost of electricity borne by the retailer and the second one charges according to a time-variant increasing-block price, where for each scheme, a zero-revenue model and a constant-rate revenue model are considered. In addition, the relationship between the studied games and certain competitive routing games from the computer networking community, known as atomic flow games, is established, for which it is shown that the proposed noncooperative game formulation falls under the class of atomic splittable flow games. The Nash equilibrium is shown to exist for four different combined cases corresponding to the two pricing schemes and the two revenue models, and is unique for three of the cases under certain conditions. It is further shown that both pricing schemes lead to similar electricity loading patterns when consumers are only interested in minimizing the electricity costs without any other profit considerations. Finally, the conditions under which the increasing-block pricing scheme is preferred over the average-cost based pricing scheme are discussed.

cs.GT↗