Search

Scanning for trojan (backdoor) in deep neural networks is crucial due to their significant real-world applications. There has been an increasing focus on developing effective general trojan scanning methods across various trojan attacks. Despite advancements, there remains a shortage of methods that perform effectively without preconceived assumptions about the backdoor attack method. Additionally, we have observed that current methods struggle to identify classifiers trojaned using adversarial training. Motivated by these challenges, our study introduces a novel scanning method named TRODO (TROjan scanning by Detection of adversarial shifts in Out-of-distribution samples). TRODO leverages the concept of "blind spots"--regions where trojaned classifiers erroneously identify out-of-distribution (OOD) samples as in-distribution (ID). We scan for these blind spots by adversarially shifting OOD samples towards in-distribution. The increased likelihood of perturbed OOD samples being classified as ID serves as a signature for trojan detection. TRODO is both trojan and label mapping agnostic, effective even against adversarially trained trojaned classifiers. It is applicable even in scenarios where training data is absent, demonstrating high accuracy and adaptability across various scenarios and datasets, highlighting its potential as a robust trojan scanning strategy., Comment: Accepted at the Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) 2024. The code repository is available at: https://github.com/rohban-lab/TRODO

Despite significant progress in Anomaly Detection (AD), the robustness of existing detection methods against adversarial attacks remains a challenge, compromising their reliability in critical real-world applications such as autonomous driving. This issue primarily arises from the AD setup, which assumes that training data is limited to a group of unlabeled normal samples, making the detectors vulnerable to adversarial anomaly samples during testing. Additionally, implementing adversarial training as a safeguard encounters difficulties, such as formulating an effective objective function without access to labels. An ideal objective function for adversarial training in AD should promote strong perturbations both within and between the normal and anomaly groups to maximize margin between normal and anomaly distribution. To address these issues, we first propose crafting a pseudo-anomaly group derived from normal group samples. Then, we demonstrate that adversarial training with contrastive loss could serve as an ideal objective function, as it creates both inter- and intra-group perturbations. However, we notice that spurious negative pairs compromise the conventional contrastive loss to achieve robust AD. Spurious negative pairs are those that should be closely mapped but are erroneously separated. These pairs introduce noise and misguide the direction of inter-group adversarial perturbations. To overcome the effect of spurious negative pairs, we define opposite pairs and adversarially pull them apart to strengthen inter-group perturbations. Experimental results demonstrate our superior performance in both clean and adversarial scenarios, with a 26.1% improvement in robust detection across various challenging benchmark datasets. The implementation of our work is available at: https://github.com/rohban-lab/COBRA., Comment: Accepted at the 13th International Conference on Learning Representations (ICLR) 2025

Localizing the seizure onset zone (SOZ) as a step of presurgical planning leads to higher efficiency in surgical and stimulation treatments. However, the clinical localization including structural, ictal, and invasive data acquisition and assessment is a difficult and long procedure with increasing challenges in patients with complex epileptic foci. The interictal methods are proposed to assist in presurgical planning with simpler data acquisition and higher speed. This study presents a spatiotemporal component classification for the localization of epileptic foci using resting-state functional magnetic resonance imaging data. This method is based on spatiotemporal independent component analysis on rsfMRI with a component-sorting procedure upon dominant power frequency, biophysical constraints, spatial lateralization, local connectivity, temporal energy, and functional non-Gaussianity. This method utilized the rs-fMRI potential to reach a high spatial accuracy in localizing epileptic foci from interictal data while retaining the reliability of results for clinical usage. Thirteen patients with temporal lobe epilepsy who underwent surgical resection and had seizure-free surgical outcomes after a 12-month follow-up were included in this study. All patients had presurgical structural MRI and rsfMRI while postsurgical MRI images were available for ten. Based on the relationship between the localized foci and resection, the results were classified into three groups fully concordant, partially concordant, and discordant. These groups had the resulting cluster aligned with, in the same lobe with, and outside the lobe of the resection area, respectively.

In this paper, We conducted an in-depth investigation into the thermodynamic topology of Einstein-Gauss-Bonnet black holes within the framework of Conformal Field Theory (CFT), considering the implications of non-extensive entropy formulations. Our study reveals that the parameter $\lambda$ (R\'{e}nyi entropy) plays a crucial role in the phase behavior of black holes. Specifically, when $\lambda$ is below the critical value (C), it has a negligible impact on the phase behavior. However, when $\lambda$ exceeds the critical value, it significantly alters the phase transition outcomes. Determining the most physically representative values of $\lambda$ will require experimental validation, but this parameter flexibility allows researchers to better explain black hole phase transitions under varying physical conditions. Furthermore, the parameters $\alpha$ and $\beta$ affect the phase structure and topological charge for the Sharma-Mittal entropy. Only in the case of $C>C_c$ and in the condition of $\alpha\approx\beta$ will we have a first-order phase transition with topological charge + 1. Additionally, for the loop quantum gravity non-extensive entropy as the parameter $q$ approaches 1, the classification of topological charges changes. We observe configurations with one and three topological charges with respect to critical value $C$, resulting in a total topological charge $W = +1$, and configurations with two topological charges $(\omega = +1, -1)$, leading to a total topological charge $W = 0$. These findings provide new insights into the complex phase behavior and topological characteristics of black holes in the context of CFT and non-extensive entropy formulations., Comment: 17 pages, 12 figures

In this work, we investigate the Heavy-Quark Spin Symmetry (HQSS) exhibited in the effective Lagrangians governing the three-point interactions of $D$ mesons, charmed baryons, and nucleons. We first construct the effective Lagrangians, and there are 12 distinct terms. As a result, we observe that the invariant Lagrangian under HQSS manifests exclusively in the pseudoscalar $D$ mesons coupling to nucleons and $\Lambda_c$ baryons, whereas nucleons and $\Sigma_c$ ($\Sigma_c^*$) baryons only couple with vector $D$ mesons. By taking into account the violated heavy-quark spin transformation, one can recover all interactions from the effective Lagrangians. Furthermore, we compute the differential cross-sections of the $p\bar p \to Y_c\bar{Y}_c'$ scatterings, where $Y_c,\bar{Y_c}' = \Lambda_c,~\Sigma_c,~\Sigma_c^*$, to reveal the residue of the violating HQSS (VHQSS) on charmed baryon production. Ultimately, by accounting for VHQSS, we aim for precise predictions of production rates, which are essential for the High-Energy Storage Ring (HESR) experiments at the Facility for Antiproton and Ion Research (FAIR)., Comment: 23 pages, 10 figures, 2 tables

We investigate a compelling model of a rotating black hole that is deformed by the effects of loop quantum gravity (LQG). We present a simplified metric and explore two distinct geometries: one in which the masses of the black hole and white hole are equal, and another in which they differ. Our analysis yields the radius of the innermost stable circular orbits (ISCO), as well as the energy and angular momentum of a particle within this framework. Additionally, we find the frequency of the first-order resonance separately. We constrain the model by the quasi-periodic oscillations (QPO) of the X-ray binary GRO J1655-40. We show that $\lambda=0.15^{+0.23}_{-0.14}$ at $1\sigma$ confidence level for equal mass black hole and white hole geometry. For the other geometry we get $\lambda=0.11^{+0.07}_{-0.07}$ at $1\sigma$ confidence level.We encounter a degeneracy in the parameter space that hinders our ability to constrain $\lambda$ with greater precision.

We explore the modifications of hadron structure in a nuclear medium, focusing on the spacelike electromagnetic form factors (EMFFs) of light and heavy-light pseudoscalar mesons. By combining the light-front quark model (LFQM) with the quark-meson coupling (QMC) model, which reasonably reproduces EMFFs in free space and the saturation properties of nuclear matter, respectively, we systematically analyze the in-medium EMFFs and charge radii of mesons with various quark flavors. Our findings show that the EMFFs of charged (neutral) mesons exhibit a faster fall-off (increase) with increasing four-momentum transfer squared and nuclear density. Consequently, the absolute value of the charge radii of mesons increases with nuclear density, where the rate of increase depends on their quark flavor contents. We observe that the EMFFs of pions and kaons undergo significant modifications in the nuclear medium, while heavy-light mesons are only slightly modified. By decomposing the quark flavor contributions to EMFFs, we show that the medium effects primarily impact the light-quark sector, leaving the heavy-quark sector nearly unaffected. The results of this study further suggest the importance of the medium effects at the quark level., Comment: 15 pages, 9 figures, 4 tables

The capabilities of automated vehicles are advancing rapidly, yet achieving full autonomy remains a significant challenge, requiring ongoing human cognition in decision-making processes. Incorporating human cognition into control algorithms has become increasingly important, as researchers work to develop strategies that minimize conflicts between human drivers and AI systems. Despite notable progress, many challenges persist, underscoring the need for further innovation and refinement in this field. This review covers recent progress in human-vehicle interaction (HVI) and AI collaboration for vehicle control. First, we start by looking at how HVI has evolved, pointing out key developments and identifying persistent problems. Second, we discuss the existing techniques, including methods for integrating human intuition and cognition into decision-making processes and developing systems that can mimic human behavior to enable optimal driving strategies and achieve safer and more efficient transportation. This review aims to contribute to the development of more effective and adaptive automated driving systems by enhancing human-AI collaboration., Comment: This is a review paper containing 10 pages and 8 figures

Ensuring software quality remains a critical challenge in complex and dynamic development environments, where software defects can result in significant operational and financial risks. This paper proposes an innovative framework for software defect prediction that combines ensemble feature extraction with reinforcement learning (RL)--based feature selection. We claim that this work is among the first in recent efforts to address this challenge at the file-level granularity. The framework extracts diverse semantic and structural features from source code using five code-specific pre-trained models. Feature selection is enhanced through a custom-defined embedding space tailored to represent feature interactions, coupled with a pheromone table mechanism inspired by Ant Colony Optimization (ACO) to guide the RL agent effectively. Using the Proximal Policy Optimization (PPO) algorithm, the proposed method dynamically identifies the most predictive features for defect detection. Experimental evaluations conducted on the PROMISE dataset highlight the framework's superior performance on the F1-Score metric, achieving an average improvement of $6.25\%$ over traditional methods and baseline models across diverse datasets. This study underscores the potential for integrating ensemble learning and RL for adaptive and scalable defect prediction in modern software systems., Comment: 14 pages 8 figures

General relativity predicts that a rotating black hole drags the spacetime due to its spin. This effect can influence the motion of nearby objects, causing them to either fall into the black hole or orbit around it. In classical Newtonian mechanics, as the radius of the orbit increases, the angular velocity of an object in a stable circular orbit decreases. However, Aschenbach discovered that for a hypothetical non-rotating observer, contrary to usual behavior, the angular velocity increases with radius in certain regions. Although the possibility of observing rare and less probable rotational behaviors in a rotating structure is not unlikely or impossible. However, observing such behaviors in a static structure is not only intriguing but also thought-provoking, as it raises questions about the factors that might play a role in such phenomena. In seeking answers to this question, various static models, particularly in the context of nonlinear fields, were examined, with some results presented as examples in the article. Among the models studied, the model of Magnetic Black Holes in 4D Einstein Gauss Bonnet Massive Gravity Coupled to Nonlinear Electrodynamics (M-EGB-Massive) appears to be a candidate for this phenomenon. In the analysis section, we will discuss the commonalities of this model with previous models that have exhibited this phenomenon and examine the cause of this phenomenon. Finally, we will state whether this phenomenon is observable in other black holes and, if not, why, Comment: 10 pages, 8 figures

Strong lensing directly probes the internal structure of the lensing galaxies. In this paper, we investigate the relation between the internal structure of massive elliptical galaxies and their environment using a sample of 15 strong lensing systems. We performed lens modeling for them using Lenstronomy and constrained the mass and light distributions of the deflector galaxies. We adopt the local galaxy density as a metric for the environment and test our results against several alternative definitions of it. We robustly find that the centroid offset between the mass and light is not correlated with the local galaxy density. This result supports using centroid offsets as a probe of dark matter theories since the environment's impact on it can be treated as negligible. Although we find a strong correlation between the position angle offset and the standard definition of the local galaxy density, consistent with previous studies, the correlation becomes weaker for alternative definitions of the local galaxy density. This result weakens the support for interpreting the position angle misalignment as having originated from interaction with the environment. Furthermore, we find the 'residual shear' magnitude in the lens model to be uncorrelated with the local galaxy density, supporting the interpretation of the residual shear originating, in part, from the inadequacy in modeling the angular structure of the lensing galaxy and not solely from the structures present in the environment or along the line of sight., Comment: 15 pages, 9 figures, 3 tables. Submitted to A&A

The integration of non-commutative geometry and Gauss-Bonnet corrections in an action and the study of their black hole responses can provide highly intriguing insights. Our primary motivation for this study is to understand the interplay of these two parameters on the geodesics of spacetime, including photon spheres and time-like orbits. In this study, we found that this integration, in its initial form, can limit the value of the Gauss-Bonnet parameter ($\alpha$), creating a critical threshold beyond which changes in the non-commutative parameter ($\Xi$) become ineffective, and the structure can only manifest as a naked singularity. Furthermore, we found that using a more complex model, which includes additional factors such as a cloud of strings and linear charge, as a sample for studying spacetime geodesics, yield different and varied results. In this scenario, negative $\alpha$ values can also play a role, notably preserving the black hole form even with a super-extremal charge ($q > m$). For $\alpha> 0.1$, the black hole mass parameter becomes significantly influential, with a critical mass below which the impact of other parameter changes is nullified. Interestingly, considering a more massive black hole, this high-mass state also maintains its black hole form within the super-extremal charge range. The existence of these two models led us to our main goal. By examining the temperature for these two cases, we find that both situations are suitable for studying the WGC. Finally, based on the behavior of these two models, we will explain how the WGC acts as a logical solution and a protector for the WCCC., Comment: 22 pages, 14 figures, Increasing the quality of writing, fixing typos, and updating references

Financial markets are integral to a country's economic success, yet their complex nature raises challenging issues for predicting their behaviors. There is a growing demand for an integrated system that explores the vast and diverse data in financial reports with powerful machine-learning models to analyze financial markets and suggest appropriate investment strategies. This research provides an end-to-end decision support system (DSS) that pervasively covers the stages of gathering, cleaning, and modeling the stock's financial and fundamental data alongside the country's macroeconomic conditions. Analyzing and modeling the fundamental data of securities is a noteworthy method that, despite its greater power, has been used by fewer researchers due to its more complex and challenging issues. By precisely analyzing securities' fundamental data, the proposed system assists investors in predicting stock future prices and allocating assets in major financial markets: stock, bond, and commodity. The most notable contributions and innovations of this research are: (1) Developing a robust predictive model for mid- to long-term stock returns, tailored for investors rather than traders, (2) The proposed DSS considers a diverse set of features relating to the economic conditions of the company, including fundamental data, stock trading characteristics, and macro-economic attributes to enhance predictive accuracy, (3) Evaluating the DSS performance on the Tehran Stock Exchange that has specific characteristics of small to medium-sized economies with high inflation rates and showing the superiority to novel researches, and (4) Empowering the DSS to generate different asset allocation strategies in various economic situations by simulating expert investor decision-making.

The CEEMDAN algorithm is one of the modern methods used in the analysis of non-stationary signals. This research presents a study of the effectiveness of this method in audio source separation to know the limits of its work. It concluded two conditions related to frequencies and amplitudes of mixed signals to be separated by CEEMDAN. The performance of the algorithm in separating noise from speech and separating speech signals from each other is studied. The research reached a conclusion that CEEMDAN can remove some types of noise from speech (speech improvement), and it cannot separate speech signals from each other (cocktail party). Simulation is done using Matlab environment and Noizeus database., Comment: in Arabic language

Non-commutative black holes(NCBH), due to the non-commutativity of spacetime coordinates, lead to a modification of the spacetime metric. By replacing the Dirac delta function with a Gaussian distribution, the mass is effectively smeared, eliminating point-like singularities. Our objective is to investigate the impact of this change on spacetime geodesics, including photon spheres and time-like orbits. We will demonstrate how the photon sphere can serve as a tool to classify spacetime, illustrating the influence of the NC parameter and constraining its values in various modes of these black holes. Additionally, using this classification, we will show how the addition of the nonlinear Einstein-Born-Infeld(BI) field to the model enhances its physical alignment with reality compared to the charged model. In the dS BI model, we will show how the study of the effective potential and photon sphere can provide insights into the initial structural status of the model, thereby establishing this potential as an effective tool for examining the initial conditions of black holes. Finally, by examining super-extremality conditions, we will show that the AdS BI model, with the necessary conditions, can be a suitable candidate for studying and observing the effects of the Weak Gravity Conjecture (WGC)., Comment: 24 pages, 15 figures

Topic modeling, or more broadly, dimensionality reduction, techniques provide powerful tools for uncovering patterns in large datasets and are widely applied across various domains. We investigate how Non-negative Matrix Factorization (NMF) can introduce bias in the representation of data groups, such as those defined by demographics or protected attributes. We present an approach, called Fairer-NMF, that seeks to minimize the maximum reconstruction loss for different groups relative to their size and intrinsic complexity. Further, we present two algorithms for solving this problem. The first is an alternating minimization (AM) scheme and the second is a multiplicative updates (MU) scheme which demonstrates a reduced computational time compared to AM while still achieving similar performance. Lastly, we present numerical experiments on synthetic and real datasets to evaluate the overall performance and trade-offs of Fairer-NMF

This paper addresses the challenge of speaker separation, which remains an active research topic despite the promising results achieved in recent years. These results, however, often degrade in real recording conditions due to the presence of noise, echo, and other interferences. This is because neural models are typically trained on synthetic datasets consisting of mixed audio signals and their corresponding ground truths, which are generated using computer software and do not fully represent the complexities of real-world recording scenarios. The lack of realistic training sets for speaker separation remains a major hurdle, as obtaining individual sounds from mixed audio signals is a nontrivial task. To address this issue, we propose a novel method for constructing a realistic training set that includes mixture signals and corresponding ground truths for each speaker. We evaluate this dataset on a deep learning model and compare it to a synthetic dataset. We got a 1.65 dB improvement in Scale Invariant Signal to Distortion Ratio (SI-SDR) for speaker separation accuracy in realistic mixing. Our findings highlight the potential of realistic training sets for enhancing the performance of speaker separation models in real-world scenarios., Comment: in Arabic language

For critical applications that require a higher level of reliability, the Triple Modular Redundancy (TMR) scheme is usually employed to implement fault-tolerant arithmetic units. However, this method imposes a significant area and power/energy overhead. Also, the majority-based voter in the typical TMR designs is highly sensitive to soft errors and the design diversity of the triplicated module, which may result in an error for a small difference between the output of the TMR modules. However, a wide range of applications deployed in critical systems are inherently error-resilient, i.e., they can tolerate some inexact results at their output while having a given level of reliability. In this paper, we propose a High Precision Redundancy Multiplier (HPR-Mul) that relies on the principles of approximate computing to achieve higher energy efficiency and lower area, as well as resolve the aforementioned challenges of the typical TMR schemes, while retaining the required level of reliability. The HPR-Mul is composed of full precision (FP) and two reduced precision (RP) multipliers, along with a simple voter to determine the output. Unlike the state-of-the-art Reduced Precision Redundancy multipliers (RPR-Mul) that require a complex voter, the voter of the proposed HPR-Mul is designed based on mathematical formulas resulting in a simpler structure. Furthermore, we use the intermediate signals of the FP multiplier as the inputs of the RP multipliers, which significantly enhance the accuracy of the HPR-Mul. The efficiency of the proposed HPR-Mul is evaluated in a 15-nm FinFET technology, where the results show up to 70% and 69% lower power consumption and area, respectively, compared to the typical TMR-based multipliers. Also, the HPR-Mul outperforms the state-of-the-art RPR-Mul by achieving up to 84% higher soft error tolerance.

In this study, we investigate the Weak Gravity Conjecture (WGC) in the context of quantum-corrected AdS-Reissner-Nordstrom (AdS-RN) black holes within Kiselev spacetime. Our focus is on photon spheres, which serve as markers for stable and unstable photon spheres. We confirm the validity of the WGC by demonstrating that quantum corrections do not alter the essential charge-to-mass ratio, thereby supporting the conjecture's universality. Our analysis reveals that black holes with a charge greater than their mass ($Q > M$) possess photon spheres or exhibit a total topological charge of the photon sphere (PS = -1), which upholds the WGC. This finding is significant as it reinforces the conjecture's applicability even in the presence of quantum corrections. Furthermore, we examine various parameter configurations to understand their impact on the WGC. Specifically, we find that configurations with $\omega = -\frac{1}{3}$ and $\omega = -1$ maintain the conjecture, indicating that these values do not disrupt the charge-to-mass ratio required by the WGC. However, for $\omega = -\frac{4}{3}$, the conjecture does not hold, suggesting that this particular parameter value leads to deviations from the expected behavior. These results open new directions for research in quantum gravity, as they highlight the importance of specific parameter values in maintaining the WGC. The findings suggest that while the WGC is robust under certain conditions, there are scenarios where it may be challenged, prompting further investigation into the underlying principles of quantum gravity, Comment: 13 pages, 6 figures, 1 Table

Preference disaggregation methods in Multi-Criteria Decision-Making (MCDM) often encounter challenges related to inconsistency and cognitive biases when deriving a value function from experts' holistic preferences. This paper introduces the Best-Worst Disaggregation (BWD) method, a novel approach that integrates the principles of the Best-Worst Method (BWM) into the disaggregation framework to enhance the consistency and reliability of derived preference models. BWD employs the "consider-the-opposite" strategy from BWM, allowing experts to provide two opposite pairwise comparison vectors of alternatives. This approach reduces cognitive load and mitigates anchoring bias, possibly leading to more reliable criteria weights and attribute value functions. An optimization model is formulated to determine the most suitable additive value function to the preferences expressed by an expert. The method also incorporates a consistency analysis to quantify and improve the reliability of the judgments. Additionally, BWD is extended to handle interval-valued preferences, enhancing its applicability in situations with uncertainty or imprecise information. We also developed an approach to identify a reference set, which is used for pairwise comparisons to elicit the value functions and weights. A case study in logistics performance evaluation demonstrates the practicality and effectiveness of BWD, showing that it produces reliable rankings aligned closely with experts' preferences., Comment: 20 pages, 4 figures

We investigate quasi-periodic oscillations (QPOs) in the context of a new rotating black hole solution that incorporates a cosmological constant. Recent work by the authors in \cite{Ovalle:2022eqb} interpreted the cosmological constant, denoted as $\Lambda$, as a form of vacuum energy and employed a gravitational decoupling approach to derive an extended Kerr-de Sitter black hole solution, which is geometrically richer than the classical case. In this study, we derive the expressions for timelike circular geodesics within this solution and, using a relativistic precision model, calculate the corresponding frequencies of the QPOs. To constrain our model, we apply Bayesian formalism, utilizing data from three well-known microquasars: GRO 1655-40, XTE 1550-564, and GRS 1915+105. Our analysis reveals that$\Lambda$ is degenerate and correlated with other parameters. Finally, we perform a Bayesian model comparison with the Kerr metric and find that the Kerr metric is favored among the models considered., Comment: Comments are welcome

In this paper, we investigate the topological charge and the conditions for the existence of the photon sphere (PS) in Kiselev-AdS black holes within \(f(R, T)\) gravity. We employ two different methods based on Duan's topological current \(\phi\)-mapping theory viz analize of temperature and the generalized Helmholtz free energy methods to study the topological classes of our black hole. By considering the mentioned black hole, we discuss the critical and zero points (topological charges and topological numbers) for different parameters. Our findings reveal that the Kiselev parameter \(\omega\) and the \(f(R, T)\) gravity parameter \(\gamma\) influence the number of topological charges of black holes, leading to novel insights into topological classifications. We observe that for given values of the free parameters, there exist total topological charges (\(Q_{total} = -1\)) for T-method and total topological numbers (\(W = +1\)) for the generalized Helmholtz free energy method. Our research findings elucidate that, in contrast to the scenario where \(\omega = 1/3\), in other cases, increasing the parameter \(\gamma\) increases the number of total topological charges for the black hole. Interestingly, for the phantom field (\(\omega = -4/3\)), we observed that decreasing the parameter \(\gamma\) increases the number of topological charges. Additionally, we study the results for the photon sphere. The studied models clearly reveal that the simultaneous presence of \(\gamma\) and \(\omega\) effectively expands the permissible range for \(\gamma\). In other words, the model can exhibit black hole behavior over a larger domain. Additionally, it is evident that with the stepwise reduction of \(\omega\), the region covered by singularity also diminishes and becomes more restricted. However, An interesting point about all three ranges is the elimination of the forbidden region in this model., Comment: 25 pages, 19 figures, 4 table

Existing 3D facial emotion modeling have been constrained by limited emotion classes and insufficient datasets. This paper introduces "Emo3D", an extensive "Text-Image-Expression dataset" spanning a wide spectrum of human emotions, each paired with images and 3D blendshapes. Leveraging Large Language Models (LLMs), we generate a diverse array of textual descriptions, facilitating the capture of a broad spectrum of emotional expressions. Using this unique dataset, we conduct a comprehensive evaluation of language-based models' fine-tuning and vision-language models like Contranstive Language Image Pretraining (CLIP) for 3D facial expression synthesis. We also introduce a new evaluation metric for this task to more directly measure the conveyed emotion. Our new evaluation metric, Emo3D, demonstrates its superiority over Mean Squared Error (MSE) metrics in assessing visual-text alignment and semantic richness in 3D facial expressions associated with human emotions. "Emo3D" has great applications in animation design, virtual reality, and emotional human-computer interaction., Comment: 11 pages, 10 figures

The cosmic birefringence (CB) angle refers to the rotation of the linear polarization plane of Cosmic Microwave Background (CMB) radiations when parity-violating theories are considered. We analyzed the Quantum Boltzmann equation for an ensemble of CMB photons interacting with the right-handed sterile neutrino dark matter (DM) and axion-like particles (ALPs) DM in the presence of the scalar metric perturbation. We used the birefringence angle of CMB to study those probable candidates of DM. It is shown that the CB angle contribution of sterile neutrino is much less that two other sources considered here. Next, we combined the results of the cosmic neutrinos' contribution and the contribution of the ALPs to producing the CMB birefringence and discussed the uncertainty on the parameter space of axions caused by the share of CMB-cosmic neutrino interaction in generating this effect. Finally, we plotted the EB power spectrum of the CMB and showed that this spectrum behaves differently in the presence of cosmic neutrinos and ALPs interactions in small $l$. Hence, future observed data for $C^{l}_{EB}$, will help us to distinguish the CB angle value due to the various sources of its production.

In this study, we investigate the properties of pseudoscalar and vector charmonia, bottomonia, and $B_c$ mesons using the light-front quark model, focusing on the $M1$ radiative transition. For that purpose, we conduct a variational analysis with a QCD-motivated effective Hamiltonian, employing a trial wave function expanded in the harmonic oscillator basis functions up to the $3S$ state. We fit the model parameters to mass spectra and decay constants, obtaining reasonable agreement with experimental data and correctly reflecting the hierarchy of mass spectra and decay constants. In analyzing the $M1$ radiative transition, we consider both good ($\mu=+$) and transverse ($\mu={\perp}$) current components with both longitudinal $(h=0)$ and transverse $(h=\pm1)$ polarizations, demonstrating that the results from both components of currents and polarizations are identical. Self-consistency is achieved by substituting $M$ with $M_0$ when computing the operators for decay constants and radiative transitions. We also find that the difference between longitudinal and transverse polarizations of the observables may quantify the anisotropy of the model wave function. Our results on radiative transitions align reasonably well with experimental data, lattice QCD, and theoretical predictions. Furthermore, we also provide predictions for $B_c$ mesons that can be tested in experiments., Comment: 24 pages, 9 figures

Unmanned aerial vehicles (UAVs) have become increasingly popular in various fields, including precision agriculture, search and rescue, and remote sensing. However, exploring unknown environments remains a significant challenge. This study aims to address this challenge by utilizing on-policy Reinforcement Learning (RL) with Proximal Policy Optimization (PPO) to explore the {two dimensional} area of interest with multiple UAVs. The UAVs will avoid collision with obstacles and each other and do the exploration in a distributed manner. The proposed solution includes actor-critic networks using deep convolutional neural networks {(CNN)} and long short-term memory (LSTM) for identifying the UAVs and areas that have already been covered. Compared to other RL techniques, such as policy gradient (PG) and asynchronous advantage actor-critic (A3C), the simulation results demonstrate the superiority of the proposed PPO approach. Also, the results show that combining LSTM with CNN in critic can improve exploration. Since the proposed exploration has to work in unknown environments, the results showed that the proposed setup can complete the coverage when we have new maps that differ from the trained maps. Finally, we showed how tuning hyper parameters may affect the overall performance.

The muon's magnetic moment is a fundamental quantity in particle physics and the deviation of its value from quantum electrodynamics (QED), motivates research beyond the standard models (SM). In this study, we utilize the effective coupling of right-handed sterile neutrinos with SM gauge bosons to calculate the muon anomalous magnetic moment ($\boldsymbol{\mu}$AMM) at one-loop level. The contribution of the sterile neutrino interactions on the $\boldsymbol{\mu}$AMM is calculated by considering the standard and non-standard neutrino interactions. Our results show that the standard sterile neutrino interactions give a negligible contribution to $\Delta a_{\boldsymbol{\mu}}$ while the non-standard neutrino interactions can play a significant role in explaining the muon $(g-2)$ anomaly. In the context of the non-standard neutrino interaction, our calculation shows that a Dirac mass scale $M_D$ around $100\,\text{GeV}$ could explain the muon anomaly if the right handed sterile neutrino's coupling with SM particles is about $\mathcal{G}_R\approx 10^{-3}$. We have also plotted the allowed region of the model parameters that satisfy the experimental data on $\Delta a_{{\boldsymbol{\mu}}}^{SN}$ and discuss the percentage of the ${\boldsymbol{\mu}}$ anomaly compensation in terms of the coupling constant $\mathcal{G}_R$.

The topology of black hole thermodynamics is a fascinating area of study that explores the connections between thermodynamic properties and topological features of black holes. This paper has led to several significant findings: We successfully derive the field equations for $F(R)$-Euler-Heisenberg theory, providing a framework for studying the interplay between modified gravity and non-linear electromagnetic effects. We obtain an analytical solution for a static, spherically symmetric, energy-dependent black hole with constant scalar curvature. Also, our analysis of black holes in F(R)-Euler-Heisenberg gravity's Rainbow reveals significant insights into their topological properties. We identified the total topological charges by examining the normalized field lines along various free parameters. Our findings indicate that the parameters $( R_0 )$ and $( f_{\epsilon} = g_{\epsilon} )$ influence the topological charges. These results are comprehensively summarized in Table I. Additionally, a general overview of Tables II, III, and IV related to the photon sphere of the mentioned black hole reveals that with an increase in $f_{\varepsilon}$, the permissible range of negative $\lambda$ in the first case gradually transitions into a non-permissible region in the third case. On the other hand, it is known that the QED parameter, which measures the strength of nonlinear effects, can be either positive or negative. A positive QED parameter reduces the electric field near the horizon and increases the black hole's mass, whereas a negative QED parameter increases the electric field and decreases the mass. According to the two statements above, it can be concluded that the increase in $f_{\varepsilon}$ actually decreases the strength of the electric field near the horizon and strengthens the effects of gravity., Comment: 21 pages, 11 figures, 4 Tables

Single-cell analysis is an increasingly relevant approach in "omics'' studies. In the last decade, it has been applied to various fields, including cancer biology, neuroscience, and, especially, developmental biology. This rise in popularity has been accompanied with creation of modern software, development of new pipelines and design of new algorithms. Many established algorithms have also been applied with varying levels of effectiveness. Currently, there is an abundance of algorithms for all steps of the general workflow. While some scientists use ready-made pipelines (such as Seurat), manual analysis is popular, too, as it allows more flexibility. Scientists who perform their own analysis face multiple options when it comes to the choice of algorithms. We have used two different datasets to test some of the most widely-used algorithms. In this paper, we are going to report the main differences between them, suggest a minimal number of algorithms for each step, and explain our suggestions. In certain stages, it is impossible to make a clear choice without further context. In these cases, we are going to explore the major possibilities, and make suggestions for each one of them., Comment: Elmi Spektr Tutoring Programme 2021, internship report

Data augmentation is important for improving machine learning model performance when faced with limited real-world data. In time series forecasting (TSF), where accurate predictions are crucial in fields like finance, healthcare, and manufacturing, traditional augmentation methods for classification tasks are insufficient to maintain temporal coherence. This research introduces two augmentation approaches using the discrete wavelet transform (DWT) to adjust frequency elements while preserving temporal dependencies in time series data. Our methods, Wavelet Masking (WaveMask) and Wavelet Mixing (WaveMix), are evaluated against established baselines across various forecasting horizons. To the best of our knowledge, this is the first study to conduct extensive experiments on multivariate time series using Discrete Wavelet Transform as an augmentation technique. Experimental results demonstrate that our techniques achieve competitive results with previous methods. We also explore cold-start forecasting using downsampled training datasets, comparing outcomes to baseline methods.

Novelty detection is a critical task for deploying machine learning models in the open world. A crucial property of novelty detection methods is universality, which can be interpreted as generalization across various distributions of training or test data. More precisely, for novelty detection, distribution shifts may occur in the training set or the test set. Shifts in the training set refer to cases where we train a novelty detector on a new dataset and expect strong transferability. Conversely, distribution shifts in the test set indicate the methods' performance when the trained model encounters a shifted test sample. We experimentally show that existing methods falter in maintaining universality, which stems from their rigid inductive biases. Motivated by this, we aim for more generalized techniques that have more adaptable inductive biases. In this context, we leverage the fact that contrastive learning provides an efficient framework to easily switch and adapt to new inductive biases through the proper choice of augmentations in forming the negative pairs. We propose a novel probabilistic auto-negative pair generation method AutoAugOOD, along with contrastive learning, to yield a universal novelty detector method. Our experiments demonstrate the superiority of our method under different distribution shifts in various image benchmark datasets. Notably, our method emerges universality in the lens of adaptability to different setups of novelty detection, including one-class, unlabeled multi-class, and labeled multi-class settings. Code: https://github.com/mojtaba-nafez/UNODE, Comment: 16 pages, 5 figures, conference

Document retrieval systems have experienced a revitalized interest with the advent of retrieval-augmented generation (RAG). RAG architecture offers a lower hallucination rate than LLM-only applications. However, the accuracy of the retrieval mechanism is known to be a bottleneck in the efficiency of these applications. A particular case of subpar retrieval performance is observed in situations where multiple documents from several different but related topics are in the corpus. We have devised a new vectorization method that takes into account the topic information of the document. The paper introduces this new method for text vectorization and evaluates it in the context of RAG. Furthermore, we discuss the challenge of evaluating RAG systems, which pertains to the case at hand., Comment: Accepted to AICT 2024

In this paper, we delve into the intricate thermodynamic topology of quantum-corrected Anti-de Sitter-Reissner-Nordstrm (AdS-RN) black hole within the framework of Kiselev spacetime. By employing the generalized off-shell Helmholtz free energy approach, we meticulously compute the thermodynamic topology of these selected black holes. Furthermore, we establish their topological classifications. Our findings reveal that quantum correction terms influence the topological charges of black holes in Kiselev spacetime, leading to novel insights into topological classifications. Our research findings elucidate that, in contrast to the scenario in which $\omega=0$ and $a=0.7$ with total topological charge $W=0$ and $\omega=-4/3$ with total topological charge $W=-1$, in other cases, the total topological charge for the black hole under consideration predominantly stabilizes at +1. This stabilization occurs with the significant influence of the parameters a, c, and $\omega$ on the number of topological charges. Specifically, when $\omega$ assumes the values of $\omega=-1/3$, $\omega=-2/3$ , $\omega=-1$, the total topological charge consistently be to W = +1., Comment: 17 pages, 5 figures. Accepted for publication in Chinese Physics C

In this paper, we delve into the study of thermodynamics and phase transition of charged Gauss-Bonnet black holes within the context of anti-de Sitter (AdS) space, with particular emphasis on the central charge's role within the dual conformal field theory (CFT). We employ a holographic methodology that interprets the cosmological constant and the Newton constant as thermodynamic variables, leading to the derivation of a modified first law of thermodynamics that incorporates the thermodynamic volume and pressure. Our findings reveal that the central charge of the CFT is intrinsically linked to the variation of these constants, and its stability can be ensured by simultaneous adjustment of these constants. We further explore the phase structures of the black holes, utilizing the free energy. Our research uncovers the existence of a critical value of the central charge, beyond which the phase diagram displays a first-order phase transition between small and large black holes. We also delve into the implications of our findings on the complexity of the CFT. Our conclusions underscore the significant role of the central charge in the holographic thermodynamics and phase transition of charged Gauss-Bonnet black holes. Furthermore, we conclude that while the central charge considered provides suitable and satisfactory solutions for this black hole in 4 and 5 dimensions, it becomes necessary to introduce a unique central charge for this structure of modified gravity. In essence, the central charge in holographic thermodynamics is not a universal value and requires modification in accordance with different modified gravities. Consequently, the physics of the problem will significantly deviate from the one discussed in this article, indicating a rich and complex landscape for future work., Comment: 26 pages, 4 figures, accepted for publication in General Relativity and Gravitation

Advancements in quantum computing underscore the critical need for sophisticated qubit readout techniques to accurately discern quantum states. This abstract presents our research intended for optimizing readout pulse fidelity for 2D and 3D Quantum Processing Units (QPUs), the latter coupled with Superconducting Radio Frequency (SRF) cavities. Focusing specifically on the application of the Least Mean Squares (LMS) adaptive filtering algorithm, we explore its integration into the FPGA-based control systems to enhance the accuracy and efficiency of qubit state detection by improving Signal-to-Noise Ratio (SNR). Implementing the LMS algorithm on the Zynq UltraScale+ RFSoC Gen 3 devices (RFSoC 4x2 FPGA and ZCU216 FPGA) using the Quantum Instrumentation Control Kit (QICK) open-source platform, we aim to dynamically test and adjust the filtering parameters in real-time to characterize and adapt to the noise profile presented in quantum computing readout signals. Our preliminary results demonstrate the LMS filter's capability to maintain high readout accuracy while efficiently managing FPGA resources. These findings are expected to contribute to developing more reliable and scalable quantum computing architectures, highlighting the pivotal role of adaptive signal processing in quantum technology advancements., Comment: Short paper submitted and accepted for QCE24 conference. 6 pages, 3 figures, 1 table, 4 equations. Paper was submitted to Quantum Technologies and Systems Engineering (QTEM) track as a New Ideas and Emergent Results (NIER) short paper

In this study, we investigate the dynamics and frame-by-frame phase transition of the first order in black hole thermodynamics. For our analysis, we will utilize the Kramers escape rate. Our focus is on charged anti-de Sitter (AdS) black holes influenced by Kaniadakis and Barrow statistics. The selection of these black holes aims to examine the effects of entropy variation on the dynamics of phase transition and to demonstrate that the Kramers escape rate, as an efficient tool, can effectively represent the dynamic transition from a small to a large black hole within the domain of first-order phase transitions. It is noteworthy that while the transition from small to large black holes should ostensibly dominate the entire process, our results indicate that the escape rate undergoes changes as it passes through the midpoint of the phase transition, leading to a reverse escape phenomenon. The findings suggest that the dynamic phase transition in charged AdS black holes affected by entropy change bears a significant resemblance to the outcomes of models influenced by Bekenstein-Hawking entropy\cite{23}. This similarity in results could serve as an additional motivation to further explore the potential capabilities of Kaniadakis and Barrow statistics in related cosmological fields. These capabilities could enhance our understanding of other cosmological properties, Comment: 20 pages, 9 figures

In this paper, we consider a K-user interference channel where interference among the users is neither too strong nor too weak, a scenario that is relatively underexplored in the literature. We propose a novel deep learning-based approach to design the encoder and decoder functions that aim to maximize the sumrate of the interference channel for discrete constellations. We first consider the MaxSINR algorithm, a state-of-the-art linear scheme for Gaussian inputs, as the baseline and then propose a modified version of the algorithm for discrete inputs. We then propose a neural network-based approach that learns a constellation mapping with the objective of maximizing the sumrate. We provide numerical results to show that the constellations learned by the neural network-based approach provide enhanced alignments, not just in beamforming directions but also in terms of the effective constellation at the receiver, thereby leading to improved sum-rate performance.

We investigate two-pion emission decays of singly charmed and bottom baryons, focusing on $\Lambda_Q^*(1P)$ and $\Xi_Q^*(1P)$ with $Q=c$ (charm) or $b$ (bottom) quarks and $J^P=1/2^-,3/2^-$, belonging to antisymmetric flavor triplet $\bar{\boldsymbol{3}}_F$. Our analysis encompasses both sequential processes, involving intermediate states belonging to symmetric flavor sextet $\boldsymbol{6}_F$ such as $\Sigma_Q(1S)$ and $\Xi_Q^\prime(1S)$ respectively with $J^P=1/2^+,3/2^+$, derived from the chiral quark model, and direct process crucial for comparison with experimental data, whose coupling constants estimated using the chiral-partner scheme. We also incorporate the convolution of the parent particle's mass for the Dalitz plot, enabling a more realistic comparison with experimental data. We scrutinize the Dalitz plots of these negative parity states in light of recent Belle measurements for $\Lambda_c(2625)^+$. Our findings support the assignment of $\Lambda_c(2625)^+$ as the $\lambda$-mode excitation with $J^P=3/2^-$ in the quark model, deduced from the the ${\Lambda_c\pi}$ invariant mass distribution, and we then give predictions for other cases, including the $\Xi_Q^*$ decays. The observed asymmetry in the ${\pi\pi}$ invariant mass distribution underscores the important role of the direct process, reflecting the chiral-partner structure in the heavy baryon sector. It is evident that the presence of the direct process is not significant in the three-body decays unless the $S$-wave resonance contribution is suppressed. We suggest further experimental verification to test our predictions and get more insights into the structure of heavy baryons., Comment: 17 pages, 11 figures

Despite recent advances in improving the sample-efficiency of reinforcement learning (RL) algorithms, designing an RL algorithm that can be practically deployed in real-world environments remains a challenge. In this paper, we present Coarse-to-fine Reinforcement Learning (CRL), a framework that trains RL agents to zoom-into a continuous action space in a coarse-to-fine manner, enabling the use of stable, sample-efficient value-based RL algorithms for fine-grained continuous control tasks. Our key idea is to train agents that output actions by iterating the procedure of (i) discretizing the continuous action space into multiple intervals and (ii) selecting the interval with the highest Q-value to further discretize at the next level. We then introduce a concrete, value-based algorithm within the CRL framework called Coarse-to-fine Q-Network (CQN). Our experiments demonstrate that CQN significantly outperforms RL and behavior cloning baselines on 20 sparsely-rewarded RLBench manipulation tasks with a modest number of environment interactions and expert demonstrations. We also show that CQN robustly learns to solve real-world manipulation tasks within a few minutes of online training., Comment: Project webpage: https://younggyo.me/cqn/

The emergence of multilingual large language models has enabled the development of language understanding and generation systems in Azerbaijani. However, most of the production-grade systems rely on cloud solutions, such as GPT-4. While there have been several attempts to develop open foundation models for Azerbaijani, these works have not found their way into common use due to a lack of systemic benchmarking. This paper encompasses several lines of work that promote open-source foundation models for Azerbaijani. We introduce (1) a large text corpus for Azerbaijani, (2) a family of encoder-only language models trained on this dataset, (3) labeled datasets for evaluating these models, and (4) extensive evaluation that covers all major open-source models with Azerbaijani support., Comment: Presented in the First Workshop on Natural Language Processing for Turkic Languages

The $N$-sum box protocol specifies a class of $\mathbb{F}_d$ linear functions $f(W_1,\cdots,W_K)=V_1W_1+V_2W_2+\cdots+V_KW_K\in\mathbb{F}_d^{m\times 1}$ that can be computed at information theoretically optimal communication cost (minimum number of qudits $\Delta_1,\cdots,\Delta_K$ sent by the transmitters Alice$_1$, Alice$_2$,$\cdots$, Alice$_K$, respectively, to the receiver, Bob, per computation instance) over a noise-free quantum multiple access channel (QMAC), when the input data streams $W_k\in\mathbb{F}_d^{m_k\times 1}, k\in[K]$, originate at the distributed transmitters, who share quantum entanglement in advance but are not otherwise allowed to communicate with each other. In prior work this set of optimally computable functions is identified in terms of a strong self-orthogonality (SSO) condition on the transfer function of the $N$-sum box. In this work we consider an `inverted' scenario, where instead of a feasible $N$-sum box transfer function, we are given an arbitrary $\mathbb{F}_d$ linear function, i.e., arbitrary matrices $V_k\in\mathbb{F}_d^{m\times m_k}$ are specified, and the goal is to characterize the set of all feasible communication cost tuples $(\Delta_1,\cdots,\Delta_K)$, not just based on $N$-sum box protocols, but across all possible quantum coding schemes. As our main result, we fully solve this problem for $K=3$ transmitters ($K\geq 4$ settings remain open). Coding schemes based on the $N$-sum box protocol (along with elementary ideas such as treating qudits as classical dits, time-sharing and batch-processing) are shown to be information theoretically optimal in all cases. As an example, in the symmetric case where rk$(V_1)$=rk$(V_2)$=rk$(V_3) \triangleq r_1$, rk$([V_1, V_2])$=rk$([V_2, V_3])$=rk$([V_3, V_1])\triangleq r_2$, and rk$([V_1, V_2, V_3])\triangleq r_3$ (rk = rank), the minimum total-download cost is $\max \{1.5r_1 + 0.75(r_3 - r_2), r_3\}$.

Deep neural networks (DNNs) are a type of artificial intelligence models that are inspired by the structure and function of the human brain, designed to process and learn from large amounts of data, making them particularly well-suited for tasks such as image and speech recognition. However, applications of DNNs are experiencing emerging growth due to the deployment of specialized accelerators such as the Google Tensor Processing Units (TPUs). In large-scale deployments, the energy efficiency of such accelerators may become a critical concern. In the voltage overscaling (VOS) technique, the operating voltage of the system is scaled down beyond the nominal operating voltage, which increases the energy efficiency and lifetime of digital circuits. The VOS technique is usually performed without changing the frequency resulting in timing errors. However, some applications such as multimedia processing, including DNNs, have intrinsic resilience against errors and noise. In this paper, we exploit the inherent resilience of DNNs to propose a quality-aware voltage overscaling framework for TPUs, named X-TPU, which offers higher energy efficiency and lifetime compared to conventional TPUs. The X-TPU framework is composed of two main parts, a modified TPU architecture that supports a runtime voltage overscaling, and a statistical error modeling-based algorithm to determine the voltage of neurons such that the output quality is retained above a given user-defined quality threshold. We synthesized a single-neuron architecture using a 15-nm FinFET technology under various operating voltage levels. Then, we extracted different statistical error models for a neuron corresponding to those voltage levels. Using these models and the proposed algorithm, we determined the appropriate voltage of each neuron. Results show that running a DNN on X-TPU can achieve 32% energy saving for only 0.6% accuracy loss.

In this paper, we investigate the weak cosmic censorship conjecture (WCCC) for the Reissner-Nordstrom (R-N) AdS black hole in a restricted phase space thermodynamics (RPST). Also here, we consider energy flux and equivalence mass-energy principle and examine the weak gravity conjecture (WGC) and the weak cosmic censorship conjecture. The incoming and outgoing energy flux leads to changes in the black hole. In that case, by applying the first law, we examined whether the second law of thermodynamics is valid. And also one can say that, in the case where absorption and superradiance are in the saturated to an equilibrium. Also, by using the thermodynamics of black holes in the restricted phase space, we show that if the black hole is in an extreme or close to an extreme state with radiation and particle absorption, the weak cosmic censorship conjecture is established. In addition, with the help of equivalence mass and energy principle and second-order approximation, in the near extremity, we find that when the black hole radiates and its central charge is greater than the scaled electric charge, the superradiance particles obey the weak gravity conjecture, and this causes the black hole to move further away from its extreme state. But when the particles that obey the weak gravity conjecture are attracted to the black hole when the black hole is very small. Then, in this case, we note that the black hole becomes closer to its extreme state., Comment: 15 pages, 1 table

The first algorithm for the Linear Quadratic (LQ) control problem with an unknown system model, featuring a regret of $\mathcal{O}(\sqrt{T})$, was introduced by Abbasi-Yadkori and Szepesv\'ari (2011). Recognizing the computational complexity of this algorithm, subsequent efforts (see Cohen et al. (2019), Mania et al. (2019), Faradonbeh et al. (2020a), and Kargin et al.(2022)) have been dedicated to proposing algorithms that are computationally tractable while preserving this order of regret. Although successful, the existing works in the literature lack a fully adaptive exploration-exploitation trade-off adjustment and require a user-defined value, which can lead to overall regret bound growth with some factors. In this work, noticing this gap, we propose the first fully adaptive algorithm that controls the number of policy updates (i.e., tunes the exploration-exploitation trade-off) and optimizes the upper-bound of regret adaptively. Our proposed algorithm builds on the SDP-based approach of Cohen et al. (2019) and relaxes its need for a horizon-dependant warm-up phase by appropriately tuning the regularization parameter and adding an adaptive input perturbation. We further show that through careful exploration-exploitation trade-off adjustment there is no need to commit to the widely-used notion of strong sequential stability, which is restrictive and can introduce complexities in initialization.

We present a novel method for accurately auditing the differential privacy (DP) guarantees of DP mechanisms. In particular, our solution is applicable to auditing DP guarantees of machine learning (ML) models. Previous auditing methods tightly capture the privacy guarantees of DP-SGD trained models in the white-box setting where the auditor has access to all intermediate models; however, the success of these methods depends on a priori information about the parametric form of the noise and the subsampling ratio used for sampling the gradients. We present a method that does not require such information and is agnostic to the randomization used for the underlying mechanism. Similarly to several previous DP auditing methods, we assume that the auditor has access to a set of independent observations from two one-dimensional distributions corresponding to outputs from two neighbouring datasets. Furthermore, our solution is based on a simple histogram-based density estimation technique to find lower bounds for the statistical distance between these distributions when measured using the hockey-stick divergence. We show that our approach also naturally generalizes the previously considered class of threshold membership inference auditing methods. We improve upon accurate auditing methods such as the $f$-DP auditing. Moreover, we address an open problem on how to accurately audit the subsampled Gaussian mechanism without any knowledge of the parameters of the underlying mechanism.

The large untapped spectrum in sub-THz allows for ultra-high throughput communication to realize many seemingly impossible applications in 6G. Phase noise (PN) is one key hardware impairment, which is accentuated as we increase the frequency and bandwidth. Furthermore, the modest output power of the power amplifier demands limits on peak to average power ratio (PAPR) signal design. In this work, we design a PN-robust, low PAPR single-carrier (SC) waveform by geometrically shaping the constellation and adapting the pulse shaping filter pair under practical PN modeling and adjacent channel leakage ratio (ACLR) constraints for a given excess bandwidth. We optimize the waveforms under conventional and state-of-the-art PN-aware demappers. Moreover, we introduce a neural-network (NN) demapper enhancing transceiver adaptability. We formulate the waveform optimization problem in its augmented Lagrangian form and use a back-propagation-inspired technique to obtain a design that is numerically robust to PN, while adhering to PAPR and ACLR constraints. The results substantiate the efficacy of the method, yielding up to 2.5 dB in the required Eb/N0 under stronger PN along with a PAPR reduction of 0.5 dB. Moreover, PAPR reductions up to 1.2 dB are possible with competitive BLER and SE performance in both low and high PN conditions., Comment: Submitted to IEEE for possible publication. arXiv admin note: text overlap with arXiv:2311.12433