Search

Facial Emotion Recognition (FER) plays a crucial role in computer vision, with significant applications in human-computer interaction, affective computing, and areas such as mental health monitoring and personalized learning environments. However, a major challenge in FER task is the class imbalance commonly found in available datasets, which can hinder both model performance and generalization. In this paper, we tackle the issue of data imbalance by incorporating synthetic data augmentation and leveraging the ResEmoteNet model to enhance the overall performance on facial emotion recognition task. We employed Stable Diffusion 2 and Stable Diffusion 3 Medium models to generate synthetic facial emotion data, augmenting the training sets of the FER2013 and RAF-DB benchmark datasets. Training ResEmoteNet with these augmented datasets resulted in substantial performance improvements, achieving accuracies of 96.47% on FER2013 and 99.23% on RAF-DB. These findings shows an absolute improvement of 16.68% in FER2013, 4.47% in RAF-DB and highlight the efficacy of synthetic data augmentation in strengthening FER models and underscore the potential of advanced generative models in FER research and applications. The source code for ResEmoteNet is available at https://github.com/ArnabKumarRoy02/ResEmoteNet, Comment: 5 pages, 4 tables, 4 figures, ICASSP 2025

We consider the interaction of a compressible fluid with a flexible plate in two space dimensions. The fluid is described by the Navier--Stokes equations in a domain that is changing in accordance with the motion of the structure. The displacement of the latter evolves according to a beam equation. Both are coupled through kinematic boundary conditions and the balance of forces. We prove that for any weak solution to the coupled system, which satisfies certain additional regularity requirements, no contact occurs between the elastic wall and the bottom of the fluid cavity. This applies to both isentropic and heat-conducting fluids. As a special case of our general theory we extend the unconditional result from Grandmont and Hillairet (Arch. Ration. Mech. Anal. 220, 1283--1333, 2016) on incompressible fluids from visco-elastic to perfectly elastic plates.

We consider the constraints on new physics affecting rare kaon and hyperon decay as parametrized by a low energy effective theory that can be obtained from high-$\rm p_T$ observables at LHC. To this end, we match the relevant low energy Wilson coefficients onto those in SMEFT at dimension six and compute their contributions to charged dilepton, monolepton, monojets, and $\rm VH$ searches by ATLAS and CMS. Constraints from rare kaon decays are generally more restrictive, with the ones arising from high-$\rm p_T$ measurements complementing them along directions in parameter space to which the former are not sensitive. We also illustrate the effect of a high luminosity LHC run in reducing the viable parameter regions., Comment: 21 pages, 11 figures

The human face is a silent communicator, expressing emotions and thoughts through its facial expressions. With the advancements in computer vision in recent years, facial emotion recognition technology has made significant strides, enabling machines to decode the intricacies of facial cues. In this work, we propose ResEmoteNet, a novel deep learning architecture for facial emotion recognition designed with the combination of Convolutional, Squeeze-Excitation (SE) and Residual Networks. The inclusion of SE block selectively focuses on the important features of the human face, enhances the feature representation and suppresses the less relevant ones. This helps in reducing the loss and enhancing the overall model performance. We also integrate the SE block with three residual blocks that help in learning more complex representation of the data through deeper layers. We evaluated ResEmoteNet on three open-source databases: FER2013, RAF-DB, and AffectNet, achieving accuracies of 79.79%, 94.76%, and 72.39%, respectively. The proposed network outperforms state-of-the-art models across all three databases. The source code for ResEmoteNet is available at https://github.com/ArnabKumarRoy02/ResEmoteNet., Comment: 5 pages, 3 figures, 3 tables

We consider the motion of a large number of heavy particles in a Newtonian fluid occupying a bounded spatial domain. When we say "heavy", we mean a particle with a mass density that approaches infinity at an appropriate rate as its radius vanishes. We show that the collective effect of heavy particles on the fluid motion is similar to the Brinkman perturbation of the Navier-Stokes system identified in the homogenization process.

Motivated by the LHCb measurement of the hyperon decay mode $\Sigma^+\to p\mu^+\mu^-$ and prospects for improvement, we revisit the estimates for the rate and muon forward-backward asymmetry within the standard model and beyond. The standard model prediction has a fourfold ambiguity, and we suggest ways to resolve it with other measurements, including possible studies of $\Sigma^+\to p e^+e^-$ in the BESIII and LHCb experiments. We use the recent BESIII measurements of $\Sigma^+\to p \gamma$ and $\Sigma^+\to N\pi$ to reduce the uncertainty in the long-distance contribution to $\Sigma^+\to p\mu^+\mu^-$. Beyond the standard model, we consider a general effective Hamiltonian at low energy with ten operators whose Wilson coefficients parametrize the new physics. We derive expressions for the $\Sigma^+\to p\mu^+\mu^-$ rate and the associated muon forward-backward asymmetry in terms of these coefficients. Finally, we implement the existing constraints from kaon decays, pointing out the extent to which this hyperon mode can constrain the directions in parameter space to which the kaons are not sensitive., Comment: 27 pages, 9 figures, text & figures improved, references added, main conclusions unchanged

We show that the collective effect of $N$ rigid bodies $(\mathcal{S}_{n,N})_{n=1}^N$ of diameters $(r_{n,N})_{n=1}^N$ immersed in an incompressible non--Newtonian fluid is negligible in the asymptotic limit $N \to \infty$ as long as their total packing volume $\sum_{n=1}^N r_{n,N}^d$, $d=2,3$ tends to zero exponentially -- $\sum_{n=1}^N r_{n,N}^d \approx A^{-N}$ -- for a certain constant $A > 1$. The result is rather surprising and in a sharp contrast with the associated homogenization problem, where the same number of obstacles can completely stop the fluid motion in the case of shear thickening viscosity. A large class of non--Newtonian fluids is included, for which the viscous stress is a subdifferential of a convex potential.

We reappraise the viability of asymmetric dark matter (ADM) realized as a Dirac fermion coupling dominantly to the Standard Model fermions. Treating the interactions of such a DM particle with quarks/leptons in an effective-interactions framework, we derive updated constraints using mono-jet searches from the Large Hadron Collider (LHC) and mono-photon searches at the Large Electron-Positron (LEP) collider. We carefully model the detectors used in these experiments, which is found to have significant impact. The constraint of efficient annihilation of the symmetric part of the ADM, as well as other observational constraints are synthesized to produce a global picture. Consistent with previous work, we find that ADM with mass in the range $1-100$ GeV is strongly constrained, thus ruling out its best motivated mass range. However, we find that leptophilic ADM remains allowed for $\gtrsim 10$ GeV DM, including bounds from colliders, direct detection, and stellar heating. We forecast that the Future Circular Collider for electron-positron collisions (FCC-ee) will improve sensitivity to DM-lepton interactions by almost an order of magnitude., Comment: v2: 29 pages, 14 figures, minor updates

For many users, a private key based wallet serves as the primary entry point to blockchains. Commonly recommended wallet authentication methods, such as mnemonics or hardware wallets, can be cumbersome. This difficulty in user onboarding has significantly hindered the adoption of blockchain-based applications. We develop zkLogin, a novel technique that leverages identity tokens issued by popular platforms (any OpenID Connect enabled platform e.g., Google, Facebook, etc.) to authenticate transactions. At the heart of zkLogin lies a signature scheme allowing the signer to sign using their existing OpenID accounts and nothing else. This improves the user experience significantly as users do not need to remember a new secret and can reuse their existing accounts. zkLogin provides strong security and privacy guarantees. Unlike prior works, zkLogin's security relies solely on the underlying platform's authentication mechanism without the need for any additional trusted parties (e.g., trusted hardware or oracles). As the name suggests, zkLogin leverages zero-knowledge proofs (ZKP) to ensure that the sensitive link between a user's off-chain and on-chain identities is hidden, even from the platform itself. zkLogin enables a number of important applications outside blockchains. It allows billions of users to produce \textit{verifiable digital content leveraging their existing digital identities}, e.g., email address. For example, a journalist can use zkLogin to sign a news article with their email address, allowing verification of the article's authorship by any party. We have implemented and deployed zkLogin on the Sui blockchain as an additional alternative to traditional digital signature-based addresses., Comment: Full version of the CCS paper

This article investigates the interaction of nematic liquid crystals modeled by a simplified Ericksen-Leslie model with a rigid body. It is shown that this problem is locally strongly well-posed, and that it also admits a unique, global strong solution for initial data close to constant equilibria. The proof of the global strong solution relies on a new splitting method for the director in a mean value zero and average part., Comment: Accepted for publication in Transactions of the American Mathematical Society

In this paper we study a mathematical model describing the movement of a colloidal particle in a fixed, bounded three dimensional container filled with a nematic liquid crystal fluid. The motion of the fluid is governed by the Beris-Edwards model for nematohydrodynamics equations, which couples the incompressible Navier-Stokes equations with a parabolic system. The dynamics of colloidal particle within the nematic liquid crystal is described by the conservation laws of linear and angular momentum. We prove the existence of global weak solutions for the coupled system.

This paper provides a framework to strong time periodic solutions of quasilinear evolution equations. The novelty of this approach is that zero is allowed to be a spectral value of the underlying linearized operator. This approach is then applied to time periodic problems associated to the Navier-Stokes equations, generalized Newtonian fluids, quasilinear reaction-diffusion systems as well as to magneto-hydrodynamics., Comment: Proof of abstract result to be fixed

In this paper, we consider the viscous, incompressible, nonlinear Boussinesq system in two and three spatial dimension. We study the existence and regularity of solutions to the Boussinesq system with nonhomogeneous boundary conditions for which the normal component for the velocity is not necessary equal to zero. We establish the existence of global weak solutions to the three-dimensional instationary nonlinear Boussinesq system but without any smallness assumption on the initial and boundary conditions.

In this paper we propose the (keyed) permutation Arion and the hash function ArionHash over $\mathbb{F}_p$ for odd and particularly large primes. The design of Arion is based on the newly introduced Generalized Triangular Dynamical System (GTDS), which provides a new algebraic framework for constructing (keyed) permutation using polynomials over a finite field. At round level Arion is the first design which is instantiated using the new GTDS. We provide extensive security analysis of our construction including algebraic cryptanalysis (e.g. interpolation and Gr\"obner basis attacks) that are particularly decisive in assessing the security of permutations and hash functions over $\mathbb{F}_p$. From an application perspective, ArionHash aims for efficient implementation in zkSNARK protocols and Zero-Knowledge proof systems. For this purpose, we exploit that CCZ-equivalence of graphs can lead to a more efficient implementation of Arithmetization-Oriented primitives. We compare the efficiency of ArionHash in R1CS and Plonk settings with other hash functions such as Poseidon, Anemoi and Griffin. For demonstrating the practical efficiency of ArionHash we implemented it with the zkSNARK libraries libsnark and Dusk Network Plonk. Our result shows that ArionHash is significantly faster than Poseidon - a hash function designed for zero-knowledge proof systems. We also found that an aggressive version of ArionHash is considerably faster than Anemoi and Griffin in a practical zkSNARK setting.

The Standard Model effective field theory (SMEFT) provides a general framework to include the dynamics of the beyond standard model physics residing at a certain higher energy scale $\Lambda$. We study the top-quark production along with a Higgs boson and a jet (tHq) at the LHC experiment within the framework of the SMEFT. First, identifying the relevant sensitive dimension-6 SMEFT operators, a strategy is developed to constrain the Wilson Coefficients (WC) corresponding to these associated SMEFT operators using the latest LHC measurements, providing a complementary way to the global-fit approach. The best-fit values of these WCs are presented. Finally, we discuss the discovery potential of the signatures of those operators. We find that a discoverable excess due to SMEFT effects can be observed in the tHq process at the LHC with the center of mass energy $\sqrt{s}=13$ TeV and integrated luminosity options ${\cal L}=$300 $\rm fb^{-1}$ and 3000 $\rm fb^{-1}$., Comment: 6 pages, 2 figures, Contribution to the International Conference on High Energy Physics (ICHEP), 2022, Bologna, Italy

Hybrid beamforming has evolved as a promising technology that offers the balance between system performance and design complexity in mmWave MIMO systems. Existing hybrid beamforming methods either impose unit-modulus constraints or a codebook constraint on the analog precoders/combiners, which in turn results in a performance-overhead tradeoff. This paper puts forth a tensor framework to handle the wideband hybrid beamforming problem, with Vandermonde constraints on the analog precoders/combiners. The proposed method strikes the balance between performance, overhead and complexity. Numerical results on a 3GPP link-level test bench reveal the efficacy of the proposed approach relative to the codebook-based method while attaining the same feedback overhead. Moreover, the proposed method is shown to achieve comparable performance to the unit-modulus approaches, with substantial reductions in overhead.

We study the prospects of a simultaneous hunt at the Large Hadron Collider (LHC) of relatively light electroweakinos and a singlet-like scalar of the $Z_3$-symmetric Next-to-Minimal Supersymmetric Standard Model ($Z_3$-NMSSM) in the cascade decays of not so heavy ($\lesssim 1.5$ TeV) top squarks that are produced in pairs at the LHC which characteristically involve the singlet-like states. We work in a scenario where the lightest (next-to-lightest) SUSY particle is bino (singlino)-like with a mass below 100 GeV ($\gtrsim 100$ GeV), whereas a pair of immediately heavier neutralinos and the lighter chargino are higgsino-like with masses in the range $\sim 500$ GeV -- 1 TeV. Further, the singlet-like scalar present in the spectrum provides a funnel for a rapid enough mutual annihilation of the LSP thus making the latter meet the experimental upper bound on its relic abundance. The scenario is motivated by its ability to offer a bino-like dark matter with such a mass unlike what is now disfavored in the MSSM while avoiding the stringent lower bounds from the LHC experiments on the masses of these involved particles and still remaining reasonably `natural'. We find that while a usual cut-based analysis (CBA) on LHC data worth 300 $\text{fb}^{-1}$ would be unable to discover such excitations, a multivariate analysis (MVA) can be reasonably sensitive to higgsino-like electroweakinos having masses $\gtrsim 650$ GeV when the lighter top squark has a mass $\gtrsim 1$ TeV. On the other hand, with 3000 $\text{fb}^{-1}$ of data these masses become accessible in a CBA while even an MVA on such a data set is unlikely to find these electroweakinos with masses around 1 TeV when the mass of the lighter top squark hits $\sim 1.5$ TeV., Comment: 38 pages, 6 figures, 6 tables, Matched with the JHEP published version

We study the top-quark production along with a Higgs boson and a jet (tHq) at the LHC experiment within the framework of the Standard Model Effective Field Theory (SMEFT). A strategy is developed to constrain the Wilson Coefficients (WC) corresponding to the associated SMEFT operators using the latest LHC measurements. The best-fit values of these WCs are presented. Finally, we demonstrate the feasibility of finding the effects of these operators on various kinematical observables of the tHq process at the LHC. We find that for a set of best-fitted values of the considered WCs, the excess of signal over the backgrounds can be achieved with a reasonable significance at the center of mass energy $\sqrt{s}=13$ TeV and for integrated luminosity options ${\cal L}=$300 $\rm fb^{-1}$ and 3000 $\rm fb^{-1}$., Comment: New results added, revision of texts, accepted for publication in JHEP

We consider a bounded domain $\Omega\subset\mathbb R^3$ and a rigid body $\mathcal{S}(t)\subset\Omega$ moving inside a viscous compressible Newtonian fluid. We exploit the roughness of the body to show that the solid collides its container in finite time. We investigate the case when the boundary of the body is of $C^{1,\alpha}$-regularity and show that collision can happen for some suitable range of $\alpha$. We also discuss some no-collision results for the smooth body case., Comment: Title changed, range on $\alpha$ improved, discussion for larger values added, no-collision result added

We consider the motion of $N$ rigid bodies -- compact sets $(\mathcal{S}^1_\varepsilon, \cdots, \mathcal{S}^N_\varepsilon )_{\varepsilon > 0}$ -- immersed in a viscous incompressible fluid contained in a domain in the Euclidean space $\mathbb{R}^d$, $d=2,3$. We show the fluid flow is not influenced by the presence of the infinitely many bodies in the asymptotic limit $\varepsilon \to 0$ and $N=N(\varepsilon)\rightarrow\infty$ as soon as \[ {\rm diam}[\mathcal{S}^i_\varepsilon ] \to 0 \ \mbox{as}\ \varepsilon \to 0 ,\ i=1,\cdots, N(\varepsilon). \] The result depends solely on the geometry of the bodies and is independent of their mass densities. Collisions are allowed and the initial data are arbitrary with finite energy.

We consider the motion of a small rigid object immersed in a viscous compressible fluid in the 3-dimensional Eucleidean space. Assuming the object is a ball of a small radius $\varepsilon$ we show that the behavior of the fluid is not influenced by the object in the asymptotic limit $\varepsilon \to 0$. The result holds for the isentropic pressure law $p(\varrho) = a \varrho^\gamma$ for any $\gamma > \frac{3}{2}$ under mild assumptions concerning the rigid body density. In particular, the latter may be bounded as soon as $\gamma > 3$. The proof uses a new method of construction of the test functions in the weak formulation of the problem, and, in particular, a new form of the so-called Bogovskii operator.