Search for: All records

Differential privacy (DP) is applied when fine-tuning pre-trained language models (LMs) to limit leakage of training examples. While most DP research has focused on improving a model’s privacy-utility tradeoff, some find that DP can be unfair to or biased against underrepresented groups. In this work, we extensively analyze the impact of DP on bias in LMs. We find differentially private training can increase the model bias against protected groups w.r.t AUC-based bias metrics. DP makes it more difficult for the model to differentiate between the positive and negative examples from the protected groups and other groups in the rest of the population. Our results also show that the impact of DP on bias is affected by both the privacy protection level and the underlying distribution of the dataset. 

As the digital world gets increasingly ingrained in our daily lives, cyberattacks—especially those involving malware—are growing more complex and common, which calls for developing innovative safeguards. Keylogger spyware, which combines keylogging and spyware functionalities, is one of the most insidious types of cyberattacks. This malicious software stealthily monitors and records user keystrokes, amassing sensitive data, such as passwords and confidential personal information, which can then be exploited. This research introduces a novel browser extension designed to effectively thwart keylogger spyware attacks. The extension is underpinned by a cutting-edge algorithm that meticulously analyzes input-related processes, promptly identifying and flagging any malicious activities. Upon detection, the extension empowers users with the immediate choice to terminate the suspicious process or validate its authenticity, thereby placing crucial real-time control in the hands of the end user. The methodology used guarantees the extension's mobility and adaptability across various platforms and devices. This paper extensively details the development of the browser extension, from its first conceptual design to its rigorous performance evaluation. The results show that the extension considerably strengthens end-user protection against cyber risks, resulting in a safer web browsing experience. The research substantiates the extension's efficacy and significant potential in reinforcing online security standards, demonstrating its ability to make web surfing safer through extensive analysis and testing. 

An investigation of high-transverse-momentum (high- $p_T$) photon-triggered jets in proton-proton ( $p\text{−}p$) and ion-ion ( $A\text{−}A$) collisions at $\sqrt{s_{NN}}=0.2$and $5.02\mathrm{TeV}$is carried out, using the multistage description of in-medium jet evolution. Monte Carlo simulations of hard scattering and energy loss in heavy-ion collisions are performed using parameters tuned in a previous study of the nuclear modification factor ( $R_{AA}$) for inclusive jets and high- $p_T$hadrons. We obtain a good reproduction of the experimental data for photon-triggered jet $R_{AA}$, as measured by the ATLAS detector, the distribution of the ratio of jet to photon $p_T$( $X_{J\gamma }$), measured by both CMS and ATLAS, and the photon-jet azimuthal correlation as measured by CMS. We obtain a moderate description of the photon-triggered jet $I_{AA}$, as measured by STAR. A noticeable improvement in the comparison is observed when one goes beyond prompt photons and includes bremsstrahlung and decay photons, revealing their significance in certain kinematic regions, particularly at $X_{J\gamma }>1$. Moreover, azimuthal angle correlations demonstrate a notable impact of bremsstrahlung photons on the distribution, emphasizing their role in accurately describing experimental results. This work highlights the success of the multistage model of jet modification to straightforwardly predict (this set of) photon-triggered jet observables. This comparison, along with the role played by bremsstrahlung photons, has important consequences on the inclusion of such observables in a future Bayesian analysis. Published by the American Physical Society2025 

The Collaboration reports a new determination of the jet transport parameter $\hat{q}$in the quark-gluon plasma (QGP) using Bayesian inference, incorporating all available inclusive hadron and jet yield suppression data measured in heavy-ion collisions at the BNL Relativistic Heavy Ion Collider (RHIC) and the CERN Large Hadron Collider (LHC). This multi-observable analysis extends the previously published Bayesian inference determination of $\hat{q}$, which was based solely on a selection of inclusive hadron suppression data. is a modular framework incorporating detailed dynamical models of QGP formation and evolution, and jet propagation and interaction in the QGP. Virtuality-dependent partonic energy loss in the QGP is modeled as a thermalized weakly coupled plasma, with parameters determined from Bayesian calibration using soft-sector observables. This Bayesian calibration of $\hat{q}$utilizes active learning, a machine-learning approach, for efficient exploitation of computing resources. The experimental data included in this analysis span a broad range in collision energy and centrality, and in transverse momentum. In order to explore the systematic dependence of the extracted parameter posterior distributions, several different calibrations are reported, based on combined jet and hadron data; on jet or hadron data separately; and on restricted kinematic or centrality ranges of the jet and hadron data. Tension is observed in comparison of these variations, providing new insights into the physics of jet transport in the QGP and its theoretical formulation. Published by the American Physical Society2025