Search

In our previous works, we defined Local Information Privacy (LIP) as a context-aware privacy notion and presented the corresponding privacy-preserving mechanism. Then we claim that the mechanism satisfies epsilon-LIP for any epsilon>0 for arbitrary Px. However, this claim is not completely correct. In this document, we provide a correction to the valid range of privacy parameters of our previously proposed LIP mechanism. Further, we propose efficient algorithms to expand the range of valid privacy parameters. Finally, we discuss the impact of updated results on our original paper's experiments, the rationale of the proposed correction and corrected results.

Deep Neural Network (DNN) based classifiers have recently been used for the modulation classification of RF signals. These classifiers have shown impressive performance gains relative to conventional methods, however, they are vulnerable to imperceptible (low-power) adversarial attacks. Some of the prominent defense approaches include adversarial training (AT) and randomized smoothing (RS). While AT increases robustness in general, it fails to provide resilience against previously unseen adaptive attacks. Other approaches, such as Randomized Smoothing (RS), which injects noise into the input, address this shortcoming by providing provable certified guarantees against arbitrary attacks, however, they tend to sacrifice accuracy. In this paper, we study the problem of designing robust DNN-based modulation classifiers that can provide provable defense against arbitrary attacks without significantly sacrificing accuracy. To this end, we first analyze the spectral content of commonly studied attacks on modulation classifiers for the benchmark RadioML dataset. We observe that spectral signatures of un-perturbed RF signals are highly localized, whereas attack signals tend to be spread out in frequency. To exploit this spectral heterogeneity, we propose Filtered Randomized Smoothing (FRS), a novel defense which combines spectral filtering together with randomized smoothing. FRS can be viewed as a strengthening of RS by leveraging the specificity (spectral Heterogeneity) inherent to the modulation classification problem. In addition to providing an approach to compute the certified accuracy of FRS, we also provide a comprehensive set of simulations on the RadioML dataset to show the effectiveness of FRS and show that it significantly outperforms existing defenses including AT and RS in terms of accuracy on both attacked and benign signals., Comment: IEEE Milcom 2024

Long-tailed problems in healthcare emerge from data imbalance due to variability in the prevalence and representation of different medical conditions, warranting the requirement of precise and dependable classification methods. Traditional loss functions such as cross-entropy and binary cross-entropy are often inadequate due to their inability to address the imbalances between the classes with high representation and the classes with low representation found in medical image datasets. We introduce a novel polynomial loss function based on Pade approximation, designed specifically to overcome the challenges associated with long-tailed classification. This approach incorporates asymmetric sampling techniques to better classify under-represented classes. We conducted extensive evaluations on three publicly available medical datasets and a proprietary medical dataset. Our implementation of the proposed loss function is open-sourced in the public repository:https://github.com/ipankhi/ALPA., Comment: 13 pages, 1 figures. Accepted in The 35th British Machine Vision Conference (BMVC24)

Low Rank Adaptation (LoRA) is a popular Parameter Efficient Fine Tuning (PEFT) method that effectively adapts large pre-trained models for downstream tasks. LoRA parameterizes model updates using low-rank matrices at each layer, significantly reducing the number of trainable parameters and, consequently, resource requirements during fine-tuning. However, the lower bound on the number of trainable parameters remains high due to the use of the low-rank matrix model. In this paper, we address this limitation by proposing a novel approach that employs a low rank tensor parametrization for model updates. The proposed low rank tensor model can significantly reduce the number of trainable parameters, while also allowing for finer-grained control over adapter size. Our experiments on Natural Language Understanding, Instruction Tuning, Preference Optimization and Protein Folding benchmarks demonstrate that our method is both efficient and effective for fine-tuning large language models, achieving a substantial reduction in the number of parameters while maintaining comparable performance.

Causal Graph Discovery (CGD) is the process of estimating the underlying probabilistic graphical model that represents joint distribution of features of a dataset. CGD-algorithms are broadly classified into two categories: (i) Constraint-based algorithms (outcome depends on conditional independence (CI) tests), (ii) Score-based algorithms (outcome depends on optimized score-function). Since, sensitive features of observational data is prone to privacy-leakage, Differential Privacy (DP) has been adopted to ensure user privacy in CGD. Adding same amount of noise in this sequential-natured estimation process affects the predictive performance of the algorithms. As initial CI tests in constraint-based algorithms and later iterations of the optimization process of score-based algorithms are crucial, they need to be more accurate, less noisy. Based on this key observation, we present CURATE (CaUsal gRaph AdapTivE privacy), a DP-CGD framework with adaptive privacy budgeting. In contrast to existing DP-CGD algorithms with uniform privacy budgeting across all iterations, CURATE allows adaptive privacy budgeting by minimizing error probability (for constraint-based), maximizing iterations of the optimization problem (for score-based) while keeping the cumulative leakage bounded. To validate our framework, we present a comprehensive set of experiments on several datasets and show that CURATE achieves higher utility compared to existing DP-CGD algorithms with less privacy-leakage.

This paper considers the $\varepsilon$-differentially private (DP) release of an approximate cumulative distribution function (CDF) of the samples in a dataset. We assume that the true (approximate) CDF is obtained after lumping the data samples into a fixed number $K$ of bins. In this work, we extend the well-known binary tree mechanism to the class of \emph{level-uniform tree-based} mechanisms and identify $\varepsilon$-DP mechanisms that have a small $\ell_2$-error. We identify optimal or close-to-optimal tree structures when either of the parameters, which are the branching factors or the privacy budgets at each tree level, are given, and when the algorithm designer is free to choose both sets of parameters. Interestingly, when we allow the branching factors to take on real values, under certain mild restrictions, the optimal level-uniform tree-based mechanism is obtained by choosing equal branching factors \emph{independent} of $K$, and equal privacy budgets at all levels. Furthermore, for selected $K$ values, we explicitly identify the optimal \emph{integer} branching factors and tree height, assuming equal privacy budgets at all levels. Finally, we describe general strategies for improving the private CDF estimates further, by combining multiple noisy estimates and by post-processing the estimates for consistency., Comment: 14 pages, 1 figure, submitted to the IEEE

GRAMS (Gamma-Ray and AntiMatter Survey) is a next-generation balloon/satellite experiment utilizing a LArTPC (Liquid Argon Time Projection Chamber), to simultaneously target astrophysical observations of cosmic MeV gamma-rays and conduct an indirect dark matter search using antimatter. While LArTPCs are widely used in particle physics experiments, they have never been operated at balloon altitudes. An engineering balloon flight with a small-scale LArTPC (eGRAMS) was conducted on July 27th, 2023, to establish a system for safely operating a LArTPC at balloon altitudes and to obtain cosmic-ray data from the LArTPC. The flight was launched from the Japan Aerospace Exploration Agency's (JAXA) Taiki Aerospace Research Field in Hokkaido, Japan. The total flight duration was 3 hours and 12 minutes, including a level flight of 44 minutes at a maximum altitude of 28.9~km. The flight system was landed on the sea and successfully recovered. The LArTPC was successfully operated throughout the flight, and about 0.5 million events of the cosmic-ray data including muons, protons, and Compton scattering gamma-ray candidates, were collected. This pioneering flight demonstrates the feasibility of operating a LArTPC in high-altitude environments, paving the way for future GRAMS missions and advancing our capabilities in MeV gamma-ray astronomy and dark matter research.

The ExploreSTEM Summer Camps 2023 were designed to deliver inclusive STEM education to students aged 14 to 22 years with disabilities. This paper presents a thorough examination of the 2023 camp program, emphasizing the pivotal role of inclusive STEM education in potentially shaping students' personal and academic trajectories. The curriculum encompassed four weeklong fundamental STEM domains: Internet of Things (IoT), Computational Engineering, Artificial Intelligence (AI), and Augmented and Virtual Reality (AR/VR). Within Camp 1, students actively engaged with Dash robots, employing dedicated programming environments to command actions and gather sensor data, fostering interactions with the IoT platform and facilitating seamless data transmission. Camp 2 was dedicated to acquainting students with foundational computational engineering principles, establishing a robust framework for comprehending intricate engineering concepts. Camp 3 commenced with insightful presentations elucidating AI applications across multifaceted industries, including engineering, healthcare, and education, illuminating AI's pervasive influence on contemporary society. The primary aim of Camp 4 was to introduce students to the immersive domains of AR and VR, showcasing their applications beyond conventional STEM disciplines into everyday life experiences. The amalgamation of informative presentations, interactive activities, and a nurturing learning environment cultivated an engaging and enriching experience for all participants. By embracing inclusivity and harnessing innovative pedagogical approaches, the ExploreSTEM Summer Camps empowered students to explore, innovate, and excel within the dynamic realm of STEM education.

The assembly of galaxy clusters is understood to be a hierarchical process with a continuous accretion of galaxies over time, which increases the cluster size and mass. Late-type galaxies that fall into clusters can undergo ram-pressure stripping, forming extended gas tails within which star formation can happen. The number, location, and tail orientations of such galaxies provide clues about the galaxy infall process, the assembly of the cluster over time, and the consequences of infall for galaxy evolution. Here, we utilise the $\sim$ 0.5 degree diameter circular field of the Ultraviolet Imaging Telescope to image six galaxy clusters at z < 0.06 that are known to contain jellyfish galaxies. We searched for stripping candidates in the ultraviolet images of these clusters, which revealed 54 candidates showing signs of unilateral extra-planar emission, due to ram-pressure stripping. Seven candidates had already been identified as likely stripping based on optical B-band imaging. We identified 47 new candidates through UV imaging. Spectroscopic redshift information is available for 39 of these candidate galaxies, of which 19 are associated with six clusters. The galaxies with spectroscopic redshifts that are not part of the clusters appear to be within structures at different redshifts identified as additional peaks in the redshift distribution of galaxies, indicating that they might be ram-pressure stripped or disturbed galaxies in other structures along the line of sight. We examine the orbital history of these galaxies based on their location in the position-velocity phase-space diagram and explore a possible connection to the orientation of the tail direction among cluster member candidates. The tails of confirmed cluster member galaxies are found to be oriented away from the cluster centre., Comment: Accepted for publication in A&A

This work contributes towards the development of an efficient and scalable open-source Secure Multi-Party Computation (SMPC) protocol on machines with moderate computational resources. We use the ABY2.0 SMPC protocol implemented on the C++ based MOTION2NX framework for secure convolutional neural network (CNN) inference application with semi-honest security. Our list of contributions are as follows. Firstly, we enhance MOTION2NX by providing a tensorized version of several primitive functions including the Hadamard product, indicator function and argmax function. Secondly, we adapt an existing Helper node algorithm, working in tandem with the ABY2.0 protocol, for efficient convolution computation to reduce execution time and RAM usage. Thirdly, we also present a novel splitting algorithm that divides the computations at each CNN layer into multiple configurable chunks. This novel splitting algorithm, providing significant reduction in RAM usage, is of independent interest and is applicable to general SMPC protocols., Comment: 20 pages, 1 figure. arXiv admin note: text overlap with arXiv:2310.10133

Recent studies have emphasized the potential of forehead-crease patterns as an alternative for face, iris, and periocular recognition, presenting contactless and convenient solutions, particularly in situations where faces are covered by surgical masks. However, collecting forehead data presents challenges, including cost and time constraints, as developing and optimizing forehead verification methods requires a substantial number of high-quality images. To tackle these challenges, the generation of synthetic biometric data has gained traction due to its ability to protect privacy while enabling effective training of deep learning-based biometric verification methods. In this paper, we present a new framework to synthesize forehead-crease image data while maintaining important features, such as uniqueness and realism. The proposed framework consists of two main modules: a Subject-Specific Generation Module (SSGM), based on an image-to-image Brownian Bridge Diffusion Model (BBDM), which learns a one-to-many mapping between image pairs to generate identity-aware synthetic forehead creases corresponding to real subjects, and a Subject-Agnostic Generation Module (SAGM), which samples new synthetic identities with assistance from the SSGM. We evaluate the diversity and realism of the generated forehead-crease images primarily using the Fr\'echet Inception Distance (FID) and the Structural Similarity Index Measure (SSIM). In addition, we assess the utility of synthetically generated forehead-crease images using a forehead-crease verification system (FHCVS). The results indicate an improvement in the verification accuracy of the FHCVS by utilizing synthetic data., Comment: Accepted at Generative AI for Futuristic Biometrics - IJCB'24 Special Session

The Programmable Liquid-crystal Active Coronagraphic Imager for the DAG telescope (PLACID) instrument is a novel high-contrast direct imaging facility that was recently delivered to the Turkish 4-m DAG telescope, with first light anticipated by the end of 2024. In a nutshell, PLACID consists in a fore-optics coronagraphic intermediate stage platform, installed in-between the TROIA XAO system and the DIRAC HAWAII-1RG focal-plane array. The PLACID project, led by a consortium of Swiss Universities contracted by the Atat\"urk University Astrophysics Research and Application Center (ATASAM), has passed the Delivery Readiness Review (DRR) milestone in September 2023, and was delivered to ATASAM campus facilities in March 2024. The PLACID commissioning activities with the calibration light source at the summit, on the DAG telescope Nasmyth platform, are foreseen to take place this fall, with first light scheduled to take place before the end of the year. When on-sky, PLACID will be the world's first ''active coronagraph'' facility, fielding a customized spatial light modulator (SLM) acting as a dynamically programmable focal-plane phase mask (FPM) coronagraph from H- to Ks-band. This will provide a wealth of novel options to observers, among which software-only abilities to change or re-align the FPM pattern in function of conditions or science requirements, free of any actuator motion. Future features will include non-common path aberrations (NCPA) self-calibration, optimized coronagraphy for binary stars, as well as coherent differential imaging (CDI). We hereby present the delivered PLACID instrument, its current capabilities, and Factory Acceptance commissioning results with relevant performance metrics., Comment: 10 pages, 9 figures, SPIE Astronomical Telescopes + Instrumentation: Ground-based and Airborne Instrumentation for Astronomy X (Yokohama 2024), Paper 130961M

The programmable Liquid-crystal Active Coronagraphic Imager for the DAG Telescope (PLACID) instrument will be installed on the Turkish 4-m Telescope by the fall of 2024 and is expected to be on-sky by the end of the year. PLACID will be the first ''active stellar coronagraph instrument'', equipped with a customized spatial light modulator (SLM), which performs as a dynamically programmable focal-plane phase mask (FPM) from H- to Ks- band. A Python-based numerical simulator of SLM-based focal-plane phase coronagraph is developed to investigate the effects of discrete pixelated FPM patterns in place of classical phase masks. The simulator currently explores the impacts of two design choices, spatial sampling in the coronagraphic focal-plane (number of SLM pixels per $\lambda$/D) and phase resolution (SLM greylevel steps). The preliminary results of the monochromatic simulations show that in ideal conditions (no wavefront errors) it is sufficient to use FPMs with spatial sampling of 10 SLM pixel per $\lambda$/D and phase resolution of 8 bits. The tool is expected to enable detailed simulations of PLACID or similar SLM-based instruments, and to help with real-time operations (optimal choice of FPM for given observing conditions) and interpretation of real data. Additionally, the tool is designed to integrate and simulate advanced operation modes, in particular focal-plane phase diversity for coherent differential imaging (CDI) of exoplanets., Comment: 6 pages, 5 figures, SPIE Astronomical Telescopes + Instrumentation: Ground-based and Airborne Instrumentation for Astronomy X (Yokohama 2024), Paper 130963E

The world's first ever ''adaptive stellar coronagraph'' facility will be the PLACID instrument, installed on Turkey's new national observatory 4-m DAG telescope. PLACID incorporates a customized spatial light modulator (SLM) acting as a dynamically addressed focal-plane phase mask (FPM) coronagraph in the H-Ks bands. This new approach to high-contrast imaging will be applied on-sky in late 2024/early 2025. We present a first estimate of the science discovery space for PLACID, in terms of known exoplanets and brown dwarfs, considering raw lab contrast, contrast ratios, limiting magnitudes, coronagraphic inner working angle etc. In the future, we will also look into predicted disk and binary or multiple stars systems imaging performance, with the latter being a possible niche science case for the instrument (adaptive FPM for multiple stars). This work will inform on the first light PLACID commissioning activities and early science on the DAG telescope and is deemed to evolve in function of future developments on the DAG AO instrumentation suite., Comment: 8 pages, 6 figures, SPIE Astronomical Telescopes + Instrumentation: Ground-based and Airborne Instrumentation for Astronomy X (Yokohama 2024), Paper 130963F

Recent advances in vision-language foundational models have enabled development of systems that can perform visual understanding and reasoning tasks. However, it is unclear if these models are robust to distribution shifts, and how their performance and generalization capabilities vary under changes in data distribution. In this project we strive to answer the question "Are vision-language foundational models robust to distribution shifts like human perception?" Specifically, we consider a diverse range of vision-language models and compare how the performance of these systems is affected by corruption based distribution shifts (such as \textit{motion blur, fog, snow, gaussian noise}) commonly found in practical real-world scenarios. We analyse the generalization capabilities qualitatively and quantitatively on zero-shot image classification task under aforementioned distribution shifts. Our code will be avaible at \url{https://github.com/shivam-chandhok/CPSC-540-Project}, Comment: UBC Report

We present the AstroSat UV Deep Field South (AUDFs), an imaging survey using the wide-field Ultraviolet Imaging Telescope on board AstroSat. AUDFs covers $\sim 236$ arcmin$^{2}$ of the sky area, including the Great Observatories Origins Deep Survey (GOODS) South field in F154W and N242W filters. The deep and shallow parts of AUDFs have exposure time $\sim 62000$ and $\sim31000$ sec respectively, in the F154W filter, while in the N242W filter, they are $\sim 64000$ and $\sim34000$ sec. These observations reached a $3\sigma$ depth of 27.2 and 27.7 AB mag with a $50\%$ completeness limit of 27 and 27.6 AB mag in the F154W and N242W filters, respectively. With the acquired depth, AUDFs is the deepest far and near-UV imaging data covering the largest area known to date at 1.2" - 1.6" spatial resolution. Two primary catalogs were constructed for the F154W and N242W filters, each containing 13495 and 19374 sources brighter than the 3$\sigma$ detection limit, respectively. Our galaxy counts power-law slope $\sim0.43$~dex~mag$^{-1}$ in the N242W filter matches well with HST/WFC3/UVIS observations. A wide range of extra-galactic science can be achieved with this unique data, such as providing a sample of galaxies emitting ionizing photons in the redshift range $z \sim 1 - 3$ and beyond; constraining the UV luminosity function, investigating the extended-UV (XUV) emission around star-forming galaxies and UV morphologies for $z < 1$. The UV catalog will enhance the legacy value of the existing optical/IR imaging and spectroscopic observations from ground and space-based telescopes on the GOODS South field., Comment: 26 pages, 19 figures, 7 tables, accepted for publication in the Astrophysical Journal Supplement Series, comments are welcome

We discover strange nonchaotic attractor (SNA) through experiments in an unforced system comprising turbulent reactive flow. While models suggest SNAs are common in dynamical systems, experimental observations are primarily limited to systems with external forcing. We observe SNA prior to the emergence of periodic oscillations from chaotic fluctuations. In complex systems, self-organization can lead to order, and inherent nonlinearity can induce chaos. The occurrence of SNA, which is nonchaotic yet nonperiodic in one such complex system, is intriguing., Comment: 14 pages, 5 figures

We investigate flame-acoustic interactions in a turbulent combustor during the state of intermittency before the onset of thermoacoustic instability using complex networks. Experiments are performed in a turbulent bluff-body stabilized dump combustor where the inlet airflow rate is varied (a) quasi-statically and (b) continuously. We construct a natural visibility graph from the local heat release rate fluctuations at each location. Comparing the average degree during epochs of high- and low-amplitude acoustic pressure oscillations (p') during the state of intermittency, we detect frequency modulation in heat release rate fluctuations. Through this approach, we discover unique spatial patterns of cross-variable coupling between the frequency of heat release rate fluctuations and amplitude of p'. The frequency of local heat release rate fluctuations increases in regions of flame anchoring owing to high-frequency excitation of the flow and flame during epochs of high-amplitude p' dynamics. On the other hand, the frequency of heat release rate fluctuations decreases in regions associated with flame front distortions by large coherent vortices. In experiments with continuously varying airflow rates, the spatial pattern of frequency modulation varies with an increase in the average amplitude of p' owing to an increase in the epochs of periodic p' dynamics and the size of vortices forming in the flow. Moreover, we discover that the frequency of heat release rate fluctuations decreases even in regions of flame anchoring when very large coherent structures are shed in the flow, causing dynamic shifts in the location of flame anchoring. Our analysis thus reveals and explains rich spatial patterns of flame-acoustic interactions in turbulent reacting flows.

Recent advances in artificial intelligence have seen Large Language Models (LLMs) demonstrate notable proficiency in causal discovery tasks. This study explores the factors influencing the performance of LLMs in causal discovery tasks. Utilizing open-source LLMs, we examine how the frequency of causal relations within their pre-training corpora affects their ability to accurately respond to causal discovery queries. Our findings reveal that a higher frequency of causal mentions correlates with better model performance, suggesting that extensive exposure to causal information during training enhances the models' causal discovery capabilities. Additionally, we investigate the impact of context on the validity of causal relations. Our results indicate that LLMs might exhibit divergent predictions for identical causal relations when presented in different contexts. This paper provides the first comprehensive analysis of how different factors contribute to LLM performance in causal discovery tasks.

We present an analysis method to search for exotic low-mass field (ELF) bursts generated during large energy astrophysical events such as supernovae, binary black hole or binary neutron star mergers, and fast radio bursts using the Global Network of Optical Magnetometers for Exotic physics searches (GNOME). In our model, the associated gravitational waves or electromagnetic signals herald the arrival of the ELF burst that interacts via coupling to the spin of fermions in the magnetometers. This enables GNOME to serve as a tool for multi-messenger astronomy. The algorithm employs a model-agnostic excess-power method to identify network-wide candidate events to be subjected to a model-dependent generalized likelihood-ratio test to determine their statistical significance. We perform the first search with this technique on GNOME data coincident with the binary black hole merger S200311bg detected by LIGO/Virgo on the 11th of March 2020 and find no significant events. We place the first lab-based limits on combinations of ELF production and coupling parameters.

Certifiable robustness gives the guarantee that small perturbations around an input to a classifier will not change the prediction. There are two approaches to provide certifiable robustness to adversarial examples: a) explicitly training classifiers with small Lipschitz constants, and b) Randomized smoothing, which adds random noise to the input to create a smooth classifier. We propose \textit{SPLITZ}, a practical and novel approach which leverages the synergistic benefits of both the above ideas into a single framework. Our main idea is to \textit{split} a classifier into two halves, constrain the Lipschitz constant of the first half, and smooth the second half via randomization. Motivation for \textit{SPLITZ} comes from the observation that many standard deep networks exhibit heterogeneity in Lipschitz constants across layers. \textit{SPLITZ} can exploit this heterogeneity while inheriting the scalability of randomized smoothing. We present a principled approach to train \textit{SPLITZ} and provide theoretical analysis to derive certified robustness guarantees during inference. We present a comprehensive comparison of robustness-accuracy tradeoffs and show that \textit{SPLITZ} consistently improves upon existing state-of-the-art approaches on MNIST and CIFAR-10 datasets. For instance, with $\ell_2$ norm perturbation budget of \textbf{$\epsilon=1$}, \textit{SPLITZ} achieves $\textbf{43.2\%}$ top-1 test accuracy on CIFAR-10 dataset compared to state-of-art top-1 test accuracy $\textbf{39.8\%}

We developed a modular silicon photomultiplier camera to detect Earth-skimming PeV to EeV tau neutrinos with the imaging atmospheric Cherenkov technique. We built two cameras, a 256-pixel camera with S14161-6050HS SiPMs for the Trinity Demonstrator located on Frisco Peak, Utah, and a 512-pixel camera with S14521-6050AN SiPMs for the EUSO-SPB2 Cherenkov Telescope. The front-end electronics are based on the eMUSIC ASIC, and the camera signals are sampled and digitized with the 100MS/s and 12-bit AGET system. Both cameras are liquid-cooled. We detail the camera concept and the results from characterizing the SiPMs, bench testing, and calibrating the two cameras., Comment: Submitted to Nuclear Instruments and Methods in Physics Research A

Planning in textual environments have been shown to be a long-standing challenge even for current models. A recent, promising line of work uses LLMs to generate a formal representation of the environment that can be solved by a symbolic planner. However, existing methods rely on a fully-observed environment where all entity states are initially known, so a one-off representation can be constructed, leading to a complete plan. In contrast, we tackle partially-observed environments where there is initially no sufficient information to plan for the end-goal. We propose PDDLEGO that iteratively construct a planning representation that can lead to a partial plan for a given sub-goal. By accomplishing the sub-goal, more information is acquired to augment the representation, eventually achieving the end-goal. We show that plans produced by few-shot PDDLEGO are 43% more efficient than generating plans end-to-end on the Coin Collector simulation, with strong performance (98%) on the more complex Cooking World simulation where end-to-end LLMs fail to generate coherent plans (4%)., Comment: In *SEM 2024

Counterfactuals, or modified inputs that lead to a different outcome, are an important tool for understanding the logic used by machine learning classifiers and how to change an undesirable classification. Even if a counterfactual changes a classifier's decision, however, it may not affect the true underlying class probabilities, i.e. the counterfactual may act like an adversarial attack and ``fool'' the classifier. We propose a new framework for creating modified inputs that change the true underlying probabilities in a beneficial way which we call Trustworthy Actionable Perturbations (TAP). This includes a novel verification procedure to ensure that TAP change the true class probabilities instead of acting adversarially. Our framework also includes new cost, reward, and goal definitions that are better suited to effectuating change in the real world. We present PAC-learnability results for our verification procedure and theoretically analyze our new method for measuring reward. We also develop a methodology for creating TAP and compare our results to those achieved by previous counterfactual methods., Comment: Accepted at the 41st International Conference on Machine Learning (ICML) 2024

With the growing adoption of machine learning (ML) systems in areas like law enforcement, criminal justice, finance, hiring, and admissions, it is increasingly critical to guarantee the fairness of decisions assisted by ML. In this paper, we study the tradeoff between fairness and accuracy under the statistical notion of equalized odds. We present a new upper bound on the accuracy (that holds for any classifier), as a function of the fairness budget. In addition, our bounds also exhibit dependence on the underlying statistics of the data, labels and the sensitive group attributes. We validate our theoretical upper bounds through empirical analysis on three real-world datasets: COMPAS, Adult, and Law School. Specifically, we compare our upper bound to the tradeoffs that are achieved by various existing fair classifiers in the literature. Our results show that achieving high accuracy subject to a low-bias could be fundamentally limited based on the statistical disparity across the groups.

This paper considers the private release of statistics of several disjoint subsets of a datasets. In particular, we consider the $\epsilon$-user-level differentially private release of sample means and variances of sample values in disjoint subsets of a dataset, in a potentially sequential manner. Traditional analysis of the privacy loss under user-level privacy due to the composition of queries to the disjoint subsets necessitates a privacy loss degradation by the total number of disjoint subsets. Our main contribution is an iterative algorithm, based on suppressing user contributions, which seeks to reduce the overall privacy loss degradation under a canonical Laplace mechanism, while not increasing the worst estimation error among the subsets. Important components of this analysis are our exact, analytical characterizations of the sensitivities and the worst-case bias errors of estimators of the sample mean and variance, which are obtained by clipping or suppressing user contributions. We test the performance of our algorithm on real-world and synthetic datasets and demonstrate improvements in the privacy loss degradation factor, for fixed estimation error. We also show improvements in the worst-case error across subsets, via a natural optimization procedure, for fixed numbers of users contributing to each subset., Comment: 43 pages, 8 figures, to be submitted to the IEEE

AMD Instinct$^\text{TM}$ MI300A is the world's first data center accelerated processing unit (APU) with memory shared between the AMD "Zen 4" EPYC$^\text{TM}$ cores and third generation CDNA$^\text{TM}$ compute units. A single memory space offers several advantages: i) it eliminates the need for data replication and costly data transfers, ii) it substantially simplifies application development and allows an incremental acceleration of applications, iii) is easy to maintain, and iv) its potential can be well realized via the abstractions in the OpenMP 5.2 standard, where the host and the device data environments can be unified in a more performant way. In this article, we provide a blueprint of the APU programming model leveraging unified memory and highlight key distinctions compared to the conventional approach with discrete GPUs. OpenFOAM, an open-source C++ library for computational fluid dynamics, is presented as a case study to emphasize the flexibility and ease of offloading a full-scale production-ready application on MI300 APUs using directive-based OpenMP programming., Comment: Accepted paper at ISC High Performance 2024

Quantum machine learning (QML) as combination of quantum computing with machine learning (ML) is a promising direction to explore, in particular due to the advances in realizing quantum computers and the hoped-for quantum advantage. A field within QML that is only little approached is quantum multi-agent reinforcement learning (QMARL), despite having shown to be potentially attractive for addressing industrial applications such as factory management, cellular access and mobility cooperation. This paper presents an aerial communication use case and introduces a hybrid quantum-classical (HQC) ML algorithm to solve it. This use case intends to increase the connectivity of flying ad-hoc networks and is solved by an HQC multi-agent proximal policy optimization algorithm in which the core of the centralized critic is replaced with a data reuploading variational quantum circuit. Results show a slight increase in performance for the quantum-enhanced solution with respect to a comparable classical algorithm, earlier reaching convergence, as well as the scalability of such a solution: an increase in the size of the ansatz, and thus also in the number of trainable parameters, leading to better outcomes. These promising results show the potential of QMARL to industrially-relevant complex use cases., Comment: Submitted to the 2024 IEEE International Conference on Quantum Computing and Engineering (QCE)

The awareness of multi-cultural human values is critical to the ability of language models (LMs) to generate safe and personalized responses. However, this awareness of LMs has been insufficiently studied, since the computer science community lacks access to the large-scale real-world data about multi-cultural values. In this paper, we present WorldValuesBench, a globally diverse, large-scale benchmark dataset for the multi-cultural value prediction task, which requires a model to generate a rating response to a value question based on demographic contexts. Our dataset is derived from an influential social science project, World Values Survey (WVS), that has collected answers to hundreds of value questions (e.g., social, economic, ethical) from 94,728 participants worldwide. We have constructed more than 20 million examples of the type "(demographic attributes, value question) $\rightarrow$ answer" from the WVS responses. We perform a case study using our dataset and show that the task is challenging for strong open and closed-source models. On merely $11.1\%$, $25.0\%$, $72.2\%$, and $75.0\%$ of the questions, Alpaca-7B, Vicuna-7B-v1.5, Mixtral-8x7B-Instruct-v0.1, and GPT-3.5 Turbo can respectively achieve $<0.2$ Wasserstein 1-distance from the human normalized answer distributions. WorldValuesBench opens up new research avenues in studying limitations and opportunities in multi-cultural value awareness of LMs., Comment: Accepted at LREC-COLING 2024. Wenlong and Debanjan contributed equally

The recent prevalence of publicly accessible, large medical imaging datasets has led to a proliferation of artificial intelligence (AI) models for cardiovascular image classification and analysis. At the same time, the potentially significant impacts of these models have motivated the development of a range of explainable AI (XAI) methods that aim to explain model predictions given certain image inputs. However, many of these methods are not developed or evaluated with domain experts, and explanations are not contextualized in terms of medical expertise or domain knowledge. In this paper, we propose a novel framework and python library, MiMICRI, that provides domain-centered counterfactual explanations of cardiovascular image classification models. MiMICRI helps users interactively select and replace segments of medical images that correspond to morphological structures. From the counterfactuals generated, users can then assess the influence of each segment on model predictions, and validate the model against known medical facts. We evaluate this library with two medical experts. Our evaluation demonstrates that a domain-centered XAI approach can enhance the interpretability of model explanations, and help experts reason about models in terms of relevant domain knowledge. However, concerns were also surfaced about the clinical plausibility of the counterfactuals generated. We conclude with a discussion on the generalizability and trustworthiness of the MiMICRI framework, as well as the implications of our findings on the development of domain-centered XAI methods for model interpretability in healthcare contexts., Comment: 14 pages, 6 figures, ACM FAccT 2024

High-Level Synthesis (HLS) Design Space Exploration (DSE) is a widely accepted approach for efficiently exploring Pareto-optimal and optimal hardware solutions during the HLS process. Several HLS benchmarks and datasets are available for the research community to evaluate their methodologies. Unfortunately, these resources are limited and may not be sufficient for complex, multi-component system-level explorations. Generating new data using existing HLS benchmarks can be cumbersome, given the expertise and time required to effectively generate data for different HLS designs and directives. As a result, synthetic data has been used in prior work to evaluate system-level HLS DSE. However, the fidelity of the synthetic data to real data is often unclear, leading to uncertainty about the quality of system-level HLS DSE. This paper proposes a novel approach, called Vaegan, that employs generative machine learning to generate synthetic data that is robust enough to support complex system-level HLS DSE experiments that would be unattainable with only the currently available data. We explore and adapt a Variational Autoencoder (VAE) and Generative Adversarial Network (GAN) for this task and evaluate our approach using state-of-the-art datasets and metrics. We compare our approach to prior works and show that Vaegan effectively generates synthetic HLS data that closely mirrors the ground truth's distribution., Comment: Accepted at Great Lakes Symposium on VLSI 2024 (GLSVLSI 24)

In this work, the problem of communicating decisions of a classifier over a noisy channel is considered. With machine learning based models being used in variety of time-sensitive applications, transmission of these decisions in a reliable and timely manner is of significant importance. To this end, we study the scenario where a probability vector (representing the decisions of a classifier) at the transmitter, needs to be transmitted over a noisy channel. Assuming that the distortion between the original probability vector and the reconstructed one at the receiver is measured via f-divergence, we study the trade-off between transmission latency and the distortion. We completely analyze this trade-off using uniform, lattice, and sparse lattice-based quantization techniques to encode the probability vector by first characterizing bit budgets for each technique given a requirement on the allowed source distortion. These bounds are then combined with results from finite-blocklength literature to provide a framework for analyzing the effects of both quantization distortion and distortion due to decoding error probability (i.e., channel effects) on the incurred transmission latency. Our results show that there is an interesting interplay between source distortion (i.e., distortion for the probability vector measured via f-divergence) and the subsequent channel encoding/decoding parameters; and indicate that a joint design of these parameters is crucial to navigate the latency-distortion tradeoff. We study the impact of changing different parameters (e.g. number of classes, SNR, source distortion) on the latency-distortion tradeoff and perform experiments on AWGN and fading channels. Our results indicate that sparse lattice-based quantization is the most effective at minimizing latency across various regimes and for sparse, high-dimensional probability vectors (i.e., high number of classes)., Comment: Submitted to IEEE Transactions on Communications

Accurately predicting the 3D human posture and the pressure exerted on the body for people resting in bed, visualized as a body mesh (3D pose & shape) with a 3D pressure map, holds significant promise for healthcare applications, particularly, in the prevention of pressure ulcers. Current methods focus on singular facets of the problem -- predicting only 2D/3D poses, generating 2D pressure images, predicting pressure only for certain body regions instead of the full body, or forming indirect approximations to the 3D pressure map. In contrast, we introduce BodyMAP, which jointly predicts the human body mesh and 3D applied pressure map across the entire human body. Our network leverages multiple visual modalities, incorporating both a depth image of a person in bed and its corresponding 2D pressure image acquired from a pressure-sensing mattress. The 3D pressure map is represented as a pressure value at each mesh vertex and thus allows for precise localization of high-pressure regions on the body. Additionally, we present BodyMAP-WS, a new formulation of pressure prediction in which we implicitly learn pressure in 3D by aligning sensed 2D pressure images with a differentiable 2D projection of the predicted 3D pressure maps. In evaluations with real-world human data, our method outperforms the current state-of-the-art technique by 25% on both body mesh and 3D applied pressure map prediction tasks for people in bed., Comment: Accepted at CVPR 2024 Project Website: https://bodymap3d.github.io/ Code: https://github.com/RCHI-Lab/BodyMAP

Idealized numerical simulation is used to explore energy sinks for lee waves trapped in their bottom-intensified generating flow. In addition to the loss to explicit dissipation and reabsorption predicted by linear wave action conservation, indirect dissipation due to a nonlinear forward cascade by parametric subharmonic instability represents a significant sink that substantially reduces reabsorption. The partition of lee-wave energy loss between reabsorption and (explicit plus indirect) dissipation is independent of subgridscale damping parameterization. Remote dissipation of freely propagating internal waves generated by shear instability at the lee-wave critical layer proves to be small. A general parameterization for lee-wave dissipation of the balanced flow requires a more complete exploration of the parameter space.