Search

The Proton Improvement Plan phase II (PIP-II) project currently under construction at FNAL will replace the existing 400 MeV normal conducting linac with a new 800 MeV superconducting linac. The beam power in the downstream rapid-cycling Booster synchrotron will be doubled by raising the machine cycle frequency from 15 to 20 Hz and by increasing the injected beam intensity by a factor 1.5. This has to be accomplished without raising uncontrolled losses beyond the administrative limit of 500 W. In addition, slip-stacking efficiency in the Recycler, the next machine in the accelerator chain, sets an upper limit on the longitudinal emittance of the beam delivered by the Booster. As part of an effort to better understand potential losses and emittance blow-up in the Booster, we have been conducting full cycle 6D simulations using the code PyORBIT. The simulations include space charge, wall impedance effects and transition crossing. In this paper, we discuss our experience with the code and present representative results for possible operational scenarios., Comment: IPAC'24 - 15th International Particle Accelerator Conference

Data poisoning attacks pose a significant threat to the integrity of machine learning models by leading to misclassification of target distribution data by injecting adversarial examples during training. Existing state-of-the-art (SoTA) defense methods suffer from limitations, such as significantly reduced generalization performance and significant overhead during training, making them impractical or limited for real-world applications. In response to this challenge, we introduce a universal data purification method that defends naturally trained classifiers from malicious white-, gray-, and black-box image poisons by applying a universal stochastic preprocessing step $\Psi_{T}(x)$, realized by iterative Langevin sampling of a convergent Energy Based Model (EBM) initialized with an image $x.$ Mid-run dynamics of $\Psi_{T}(x)$ purify poison information with minimal impact on features important to the generalization of a classifier network. We show that EBMs remain universal purifiers, even in the presence of poisoned EBM training data, and achieve SoTA defense on leading triggered and triggerless poisons. This work is a subset of a larger framework introduced in \pgen with a more detailed focus on EBM purification and poison defense., Comment: arXiv admin note: substantial text overlap with arXiv:2405.18627

With the advent of AI technologies, humans and robots are increasingly teaming up to perform collaborative tasks. To enable smooth and effective collaboration, the topic of value alignment (operationalized herein as the degree of dynamic goal alignment within a task) between the robot and the human is gaining increasing research attention. Prior literature on value alignment makes an inherent assumption that aligning the values of the robot with that of the human benefits the team. This assumption, however, has not been empirically verified. Moreover, prior literature does not account for human's trust in the robot when analyzing human-robot value alignment. Thus, a research gap needs to be bridged by answering two questions: How does alignment of values affect trust? Is it always beneficial to align the robot's values with that of the human? We present a simulation study and a human-subject study to answer these questions. Results from the simulation study show that alignment of values is important for trust when the overall risk level of the task is high. We also present an adaptive strategy for the robot that uses Inverse Reinforcement Learning (IRL) to match the values of the robot with those of the human during interaction. Our simulations suggest that such an adaptive strategy is able to maintain trust across the full spectrum of human values. We also present results from an empirical study that validate these findings from simulation. Results indicate that real-time personalized value alignment is beneficial to trust and perceived performance by the human when the robot does not have a good prior on the human's values., Comment: This is a preprint of the following chapter: Bhat et al., Value Alignment and Trust in Human-Robot Interaction: Insights from Simulation and User Study, published in "Emerging Frontiers in Human-Robot Interaction", edited by Ramana Kumar Vinjamuri, 2024, Springer Nature reproduced with permission of Springer Nature. The final authenticated version is available online at: [INSERT LINK HERE]

Since molecules are ubiquitous in space, the study of the 'Molecular Universe' could unfold the mystery of the existing Interstellar medium. Star formation is linked to the chemical evolution processes. Thus, an analysis of the formation of stars coupled with the chemical evolution would give a clear insight into the entire process. For example, various evolutionary stages of star formation could be probed by observing various molecules. Chemical diagnostics of these regions could be used to extract the physical properties (e.g., density, temperature, ionization degree, etc.) of these regions. Radiative transfer calculations are worthwhile in estimating physical parameters of the region where molecules are detected. However, the radiative transfer calculations are limited due to insufficient molecular data, such as spectroscopic information or collisional excitation probabilities of many interstellar species. Complex organic molecules are detected in various environments ranging from the cold gas in prestellar cores to the warm gas on solar system scales close to individual protostars. A comparative study of the relative abundances of molecules could provide insights into the beginning of chemical complexity and the link to our solar system. In my thesis, I would mainly investigate the physical properties and kinematics of different star-forming regions using radiative transfer modeling. The observed spatial differentiation between various key molecules is used to explain their physical structure or evolution and various microphysical effects. In addition, some key molecules are used to study the various evolutionary phases. This simulated data is useful for interpreting the observed data of different telescopes like IRAM 30m, GBT, ALMA, Herschel, SOFIA, etc., Comment: PhD thesis

Train-time data poisoning attacks threaten machine learning models by introducing adversarial examples during training, leading to misclassification. Current defense methods often reduce generalization performance, are attack-specific, and impose significant training overhead. To address this, we introduce a set of universal data purification methods using a stochastic transform, $\Psi(x)$, realized via iterative Langevin dynamics of Energy-Based Models (EBMs), Denoising Diffusion Probabilistic Models (DDPMs), or both. These approaches purify poisoned data with minimal impact on classifier generalization. Our specially trained EBMs and DDPMs provide state-of-the-art defense against various attacks (including Narcissus, Bullseye Polytope, Gradient Matching) on CIFAR-10, Tiny-ImageNet, and CINIC-10, without needing attack or classifier-specific information. We discuss performance trade-offs and show that our methods remain highly effective even with poisoned or distributionally shifted generative model training data.

The phenomenon of pulsar nulling, where pulsars temporarily and stochastically cease their radio emission, is thought to be indicative of a `dying' pulsar, where radio emission ceases entirely. Here we report the discovery of a long-period pulsar, PSR J0452-3418, from the ongoing Southern-sky MWA Rapid Two-meter (SMART) pulsar survey. The pulsar has a rotation period of ${\sim}$1.67\,s and a dispersion measure of 19.8\,\dmu, and it exhibits both quasi-periodic nulling and sub-pulse drifting. Periodic nulling is uncommon, only reported in $<1$\% of the pulsar population, with even a smaller fraction showing periodic nulling and sub-pulse drifting. We describe the discovery and follow-up of the pulsar, including a positional determination using high-resolution imaging with the upgraded Giant Metrewave Radio Telescope (uGMRT), initial timing analysis using the combination of MWA and uGMRT data, and detailed characterisation of the nulling and drifting properties in the MWA's frequency band (140-170\,MHz). Our analysis suggests a nulling fraction of 34$\pm6$\% and a nulling periodicity of 42$^{+1.5}_{-1.3}$ pulses. We measure the phase ($P_2$) and time modulation ($P_3$) caused by the sub-pulse drifting, with an average $P_2$ of 7.1$^{+26.3}_{-3.1}$ degrees and a $P_3$ of 4.8$^{+1.5}_{-0.9}$ pulses. We compare and contrast the observed properties with those of other pulsars that exhibit sub-pulse drifting and quasi-periodic nulling phenomena, and find that the majority of these objects tend to be in the `death valley' in the period-period derivative ($P$-$\dot{P}$) diagram. We also discuss some broader implications for pulsar emission physics and the detectability of similar objects using next-generation pulsar surveys., Comment: 16 pages, 9 Figures, 4 Tables, Accepted for ApJ***

We present a two-dimensional model of a Fermionic wire which shows a power-law conductance behavior despite the presence of uncorrelated disorder along the direction of the transport. The power-law behavior is attributed to the presence of energy eigenstates of diverging localization length below some energy cutoff, $E_c$. To study transport, we place the wire in contact with electron reservoirs biased around a Fermi level, $E$. We show that the conductance scales super-diffusively for $|E|E_c$. At $|E|=E_c$, we show that the conductance scales diffusively or with different sub-diffusive power-laws depending on the sign of the expectation value of the disorder and the parameters of the wire., Comment: 10 pages, 4 figures

The notions of joint and outer spectral radii are extended to the setting of Hilbert $C^*$-bimodules. A Rota-Strang type characterisation is proved for the joint spectral radius. In this general setting, an approximation result for the joint spectral radius in terms of the outer spectral radius has been established. This work leads to a new proof of the Wielandt-Friedland's formula for the spectral radius of positive maps. Following an idea of J. E. Pascoe, a positive map called the maximal part has been associated to any positive map with non-zero spectral radius, on finite dimensional $C^*$-algebras. This provides a constructive treatment of the Perron-Frobenius theorem. It is seen that the maximal part of a completely positive map has a very simple structure and it is irreducible if and only if the original map is irreducible. It is observed that algebras generated by tuples of matrices can be determined and their dimensions can be computed by realizing them as linear span of Choi-Kraus coefficients of some easily computable completely positive maps., Comment: 35 pages

Continual learning (CL) remains a significant challenge for deep neural networks, as it is prone to forgetting previously acquired knowledge. Several approaches have been proposed in the literature, such as experience rehearsal, regularization, and parameter isolation, to address this problem. Although almost zero forgetting can be achieved in task-incremental learning, class-incremental learning remains highly challenging due to the problem of inter-task class separation. Limited access to previous task data makes it difficult to discriminate between classes of current and previous tasks. To address this issue, we propose `Attention-Guided Incremental Learning' (AGILE), a novel rehearsal-based CL approach that incorporates compact task attention to effectively reduce interference between tasks. AGILE utilizes lightweight, learnable task projection vectors to transform the latent representations of a shared task attention module toward task distribution. Through extensive empirical evaluation, we show that AGILE significantly improves generalization performance by mitigating task interference and outperforming rehearsal-based approaches in several CL scenarios. Furthermore, AGILE can scale well to a large number of tasks with minimal overhead while remaining well-calibrated with reduced task-recency bias., Comment: Published at 3rd Conference on Lifelong Learning Agents (CoLLAs 2024)

Fast Radio Bursts (FRBs) are millisecond dispersed radio pulses of predominately extra-galactic origin. Although originally discovered at GHz frequencies, most FRBs have been detected between 400 to 800 MHz. Nevertheless, only a handful of FRBs were detected below 400 MHz. Searching for FRBs at low frequencies is computationally challenging due to increased dispersive delay that must be accounted for. However, the wide field of view (FoV) of low-frequency telescopes - such as the the Murchison Widefield Array (MWA), and prototype stations of the low-frequency Square Kilometre Array (SKA-Low) - makes them promising instruments to open a low-frequency window on FRB event rates, and constrain FRB emission models. The standard approach, inherited from high-frequencies, is to form multiple tied-array beams to tessellate the entire FoV and perform the search on the resulting time series. This approach, however, may not be optimal for low-frequency interferometers due to their large FoVs and high spatial resolutions leading to a large number of beams. Consequently, there are regions of parameter space in terms of number of antennas and resolution elements (pixels) where interferometric imaging is computationally more efficient. Here we present a new high-time resolution imager BLINK implemented on modern Graphical Processing Units (GPUs) and intended for radio astronomy. The main goal for this imager is to become part of a fully GPU-accelerated FRB search pipeline. We describe the imager and present its verification on real and simulated data processed to form all-sky and widefield images from the MWA and prototype SKA-Low stations. We also present and compare benchmarks of the GPU and CPU code executed on laptops, desktop computers, and Australian supercomputers. The code is publicly available at https://github.com/PaCER-BLINK-Project/imager and can be applied to data from any radio telescope., Comment: 19 pages, 9 figures, 6 tables. Accepted for publication in PASA

We present a new technique that explains the output of a CNN-based model using a combination of GradCAM and LRP methods. Both of these methods produce visual explanations by highlighting input regions that are important for predictions. In the new method, the explanation produced by GradCAM is first processed to remove noises. The processed output is then multiplied elementwise with the output of LRP. Finally, a Gaussian blur is applied on the product. We compared the proposed method with GradCAM and LRP on the metrics of Faithfulness, Robustness, Complexity, Localisation and Randomisation. It was observed that this method performs better on Complexity than both GradCAM and LRP and is better than atleast one of them in the other metrics.

This paper explores the dual space corresponding to p-Bergman space and examines the essential condition for the dual space to be a q-Bergman space. The investigation involves a detailed examination of the interpolation space of a Banach couple. Additionally, we draw comparisons between the `duality', `integrability' and `regularity' properties of a domain.

GW230529 is the first compact binary coalescence detected by the LIGO-Virgo-KAGRA collaboration with at least one component mass confidently in the lower mass-gap, corresponding to the range 3-5$M_{\odot}$. If interpreted as a neutron star-black hole merger, this event has the most symmetric mass ratio detected so far and therefore has a relatively high probability of producing electromagnetic (EM) emission. However, no EM counterpart has been reported. At the merger time $t_0$, Swift-BAT and Fermi-GBM together covered 100$\%$ of the sky. Performing a targeted search in a time window $[t_0-20 \text{s},t_0+20 \text{s}]$, we report no detection by the Swift-BAT and the Fermi-GBM instruments. Combining the position-dependent $\gamma-$ray flux upper limits and the gravitational-wave posterior distribution of luminosity distance, sky localization and inclination angle of the binary, we derive constraints on the characteristic luminosity and structure of the jet possibly launched during the merger. Assuming a top-hat jet structure, we exclude at 90$\%$ credibility the presence of a jet which has at the same time an on-axis isotropic luminosity $\gtrsim 10^{48}$ erg s$^{-1}$, in the bolometric band 1 keV-10 MeV, and a jet opening angle $\gtrsim 15$ deg. Similar constraints are derived testing other assumptions about the jet structure profile. Excluding GRB 170817A, the luminosity upper limits derived here are below the luminosity of any GRB observed so far., Comment: 18 pages, 1 table, 11 figures