Search

This is a report of an NSF workshop titled "Envisioning National Resources for Artificial Intelligence Research" held in Alexandria, Virginia, in May 2024. The workshop aimed to identify initial challenges and opportunities for national resources for AI research (e.g., compute, data, models, etc.) and to facilitate planning for the envisioned National AI Research Resource. Participants included AI and cyberinfrastructure (CI) experts. The report outlines significant findings and identifies needs and recommendations from the workshop.

Text-to-video generation enhances content creation but is highly computationally intensive: The computational cost of Diffusion Transformers (DiTs) scales quadratically in the number of pixels. This makes minute-length video generation extremely expensive, limiting most existing models to generating videos of only 10-20 seconds length. We propose a Linear-complexity text-to-video Generation (LinGen) framework whose cost scales linearly in the number of pixels. For the first time, LinGen enables high-resolution minute-length video generation on a single GPU without compromising quality. It replaces the computationally-dominant and quadratic-complexity block, self-attention, with a linear-complexity block called MATE, which consists of an MA-branch and a TE-branch. The MA-branch targets short-to-long-range correlations, combining a bidirectional Mamba2 block with our token rearrangement method, Rotary Major Scan, and our review tokens developed for long video generation. The TE-branch is a novel TEmporal Swin Attention block that focuses on temporal correlations between adjacent tokens and medium-range tokens. The MATE block addresses the adjacency preservation issue of Mamba and improves the consistency of generated videos significantly. Experimental results show that LinGen outperforms DiT (with a 75.6% win rate) in video quality with up to 15$\times$ (11.5$\times$) FLOPs (latency) reduction. Furthermore, both automatic metrics and human evaluation demonstrate our LinGen-4B yields comparable video quality to state-of-the-art models (with a 50.5%, 52.1%, 49.1% win rate with respect to Gen-3, LumaLabs, and Kling, respectively). This paves the way to hour-length movie generation and real-time interactive video generation. We provide 68s video generation results and more examples in our project website: https://lineargen.github.io/., Comment: 20 pages, 20 figures

This paper describes a semi-automatic pipeline to generate challenging question-answer-decoy sets for understanding long videos. Many existing video datasets and models are focused on short clips (10s-30s). While some long video datasets do exist, they can often be solved by powerful image models applied per frame (and often to very few frames) in a video, and are usually manually annotated at high cost. In order to mitigate both these problems, we propose a scalable dataset creation pipeline which leverages large models (VLMs and LLMs), to automatically generate dense, time-aligned video captions, as well as tough question answer decoy sets for video segments (up to 15 minutes in length). Our dataset Neptune covers a broad range of long video reasoning abilities and consists of a subset that emphasizes multimodal reasoning. Since existing metrics for open-ended question answering are either rule-based or may rely on proprietary models, we provide a new open source model-based metric GEM to score open-ended responses on Neptune. Benchmark evaluations reveal that most current open-source long video models perform poorly on Neptune, particularly on questions testing temporal ordering, counting and state changes. Through Neptune, we aim to spur the development of more advanced models capable of understanding long videos. The dataset is available at https://github.com/google-deepmind/neptune

This paper tackles the intricate challenge of video question-answering (VideoQA). Despite notable progress, current methods fall short of effectively integrating questions with video frames and semantic object-level abstractions to create question-aware video representations. We introduce Local-Global Question Aware Video Embedding (LGQAVE), which incorporates three major innovations to integrate multi-modal knowledge better and emphasize semantic visual concepts relevant to specific questions. LGQAVE moves beyond traditional ad-hoc frame sampling by utilizing a cross-attention mechanism that precisely identifies the most relevant frames concerning the questions. It captures the dynamics of objects within these frames using distinct graphs, grounding them in question semantics with the miniGPT model. These graphs are processed by a question-aware dynamic graph transformer (Q-DGT), which refines the outputs to develop nuanced global and local video representations. An additional cross-attention module integrates these local and global embeddings to generate the final video embeddings, which a language model uses to generate answers. Extensive evaluations across multiple benchmarks demonstrate that LGQAVE significantly outperforms existing models in delivering accurate multi-choice and open-ended answers.

Polyp segmentation in colonoscopy is crucial for detecting colorectal cancer. However, it is challenging due to variations in the structure, color, and size of polyps, as well as the lack of clear boundaries with surrounding tissues. Traditional segmentation models based on Convolutional Neural Networks (CNNs) struggle to capture detailed patterns and global context, limiting their performance. Vision Transformer (ViT)-based models address some of these issues but have difficulties in capturing local context and lack strong zero-shot generalization. To this end, we propose the Mamba-guided Segment Anything Model (SAM-Mamba) for efficient polyp segmentation. Our approach introduces a Mamba-Prior module in the encoder to bridge the gap between the general pre-trained representation of SAM and polyp-relevant trivial clues. It injects salient cues of polyp images into the SAM image encoder as a domain prior while capturing global dependencies at various scales, leading to more accurate segmentation results. Extensive experiments on five benchmark datasets show that SAM-Mamba outperforms traditional CNN, ViT, and Adapter-based models in both quantitative and qualitative measures. Additionally, SAM-Mamba demonstrates excellent adaptability to unseen datasets, making it highly suitable for real-time clinical use.

The linear-quadratic Generalized uncertainty principle (LQG) is consistent with predictions of a minimum measurable length and a maximum measurable momentum put forth by various theories of quantum gravity. The quantum gravity effect is incorporated into a black hole (BH) by modifying its ADM mass. In this article, we explore the impact of GUP on the optical properties of an LQG modified \k BH (LQKBH). We analyze the horizon structure of the BH, which reveals a critical spin value of $7M/8$. BHs with spin $(a)$ less than the critical value are possible for any real GUP parameter $\a$ value. However, as the spin increases beyond the critical value, a forbidden region in $\a$ values pops up that disallows the existence of BHs. This forbidden region widens as we increase the spin. We then examine the impact of $\a$ on the shape and size of the BH shadow for inclination angles $17^o$ and $90^o$, providing a deeper insight into the unified effect of spin and GUP on the shadow. The size of the shadow has a minimum at $\a=1.0M$, whereas, for the exact value of $\a$, the deviation of the shadow from circularity becomes maximum when the spin is less than the critical value. No extrema is observed for $a\,>\, 7M/8$. The shadow's size and deviation are adversely affected by a decrease in the inclination angle. Finally, we confront theoretical predictions with observational results for supermassive BHs $M87^*$ and $SgrA^*$ provided by the EHT collaboration to extract bounds on the spin $a$ and GUP parameter $\a$. We explore bounds on the angular diameter $\th_d$, axial ratio $D_x$, and the deviation from \s radius $\d$ for constructing constraints on $a$ and $\a$. Our work makes LQKBHs plausible candidates for astrophysical BHs.

We obtain a totient analogue for Linnik's theorem in arithmetic progressions. Specifically, for any coprime pair of positive integers $(m,a)$ such that $m$ is odd, there exists $n\le m^{2+o(1)}$ such that $\varphi(n)\equiv a\,\mathrm{mod}\,{m}$.

We develop task scaling laws and model ladders to predict the individual task performance of pretrained language models (LMs) in the overtrained setting. Standard power laws for language modeling loss cannot accurately model task performance. Therefore, we leverage a two-step prediction approach: first use model and data size to predict a task-specific loss, and then use this task loss to predict task performance. We train a set of small-scale "ladder" models, collect data points to fit the parameterized functions of the two prediction steps, and make predictions for two target models: a 7B model trained to 4T tokens and a 13B model trained to 5T tokens. Training the ladder models only costs 1% of the compute used for the target models. On four multiple-choice tasks written in ranked classification format, we can predict the accuracy of both target models within 2 points of absolute error. We have higher prediction error on four other tasks (average absolute error 6.9) and find that these are often tasks with higher variance in task metrics. We also find that using less compute to train fewer ladder models tends to deteriorate predictions. Finally, we empirically show that our design choices and the two-step approach lead to superior performance in establishing scaling laws.

We show that for $5/6$-th of all primes $p$, Hilbert's 10-th Problem is unsolvable for $\mathbb{Q}(\zeta_3, \sqrt[3]{p})$. We also show that there is an infinite set $S$ of square free integers such tha Hilbert's 10-th Problem is unsolvable over the number fields $\mathbb{Q}(\zeta_3, \sqrt{D}, \sqrt[3]{p})$ for every $D \in S$ and every prime $p \equiv 2,5 \pmod{9}$. We use the CM elliptic curves $Y^2=X^3-432D^2$ associated to the cube sum problem, with $D$ varying in suitable congruence class, in our proof.

The Unruh effect is a well-understood phenomenon, where one considers a vacuum state of a quantum field in Minkowski spacetime, which appears to be thermally populated for a uniformly accelerating Rindler observer. In this article, we derive a variant of the Unruh effect involving two distinct accelerating observers and aim to address the following questions: (i) Is it possible to selectively thermalize a subset of momentum modes for the case of massless scalar fields, and (ii) Is it possible to excite only the left-handed massless fermions while keeping right-handed fermions in a vacuum state or vice versa? To this end, we consider a Rindler wedge $R_1$ constructed from a class of accelerating observers and another Rindler wedge $R_2$ (with $R_2 \subset R_1$) constructed from another class of accelerating observers such that the wedge $R_2$ is displaced along a null direction w.r.t $R_1$ by a parameter $\Delta$. By first considering a massless scalar field in the $R_1$ vacuum, we show that if we choose the displacement $\Delta$ along one null direction, the positive momentum modes are thermalized, whereas negative momentum modes remain in vacuum (and vice versa if we choose the displacement along the other null direction). We then consider a massless fermionic field in a vacuum state in $R_1$ and show that the reduced state in $R_2$ is such that the left-handed fermions are excited and are thermal for large frequencies. In contrast, the right-handed fermions have negligible particle density and vice versa. We argue that the toy models involving shifted Rindler spacetime may provide insights into the particle excitation aspects of evolving horizons and the possibility of Rindler spacetime having a quantum strand of hair. Additionally, based on our work, we hypothesize that massless fermions underwent selective chiral excitations during the radiation-dominated era of cosmology., Comment: 17 pages, 4 figures

Private inference (PI) serves an important role in guaranteeing the privacy of user data when interfacing with proprietary machine learning models such as LLMs. However, PI remains practically intractable due to the massive latency costs associated with nonlinear functions present in LLMs. Existing works have focused on improving latency of specific LLM nonlinearities (such as the Softmax, or the GeLU) via approximations. However, new types of nonlinearities are regularly introduced with new LLM architectures, and this has led to a constant game of catch-up where PI researchers attempt to optimize the newest nonlinear function. We introduce TruncFormer, a framework for taking any LLM and transforming it into a plaintext emulation of PI. Our framework leverages the fact that nonlinearities in LLMs are differentiable and can be accurately approximated with a sequence of additions, multiplications, and truncations. Further, we decouple the add/multiply and truncation operations, and statically determine where truncations should be inserted based on a given field size and input representation size. This leads to latency improvements over existing cryptographic protocols that enforce truncation after every multiplication operation. We open source our code for community use.

We investigate the quantum speed limit (QSL) during the time evolution of neutrino-antineutrino system under the influence of the gravitational field of a spinning primordial black hole (PBH). We derive an analytical expression for the four-vector gravitational potential in the underlying Hermitian Dirac Hamiltonian using the Boyer-Lindquist (BL) coordinates. This gravitational potential leads to an axial vector term in the Dirac equation in curved spacetime, contributing to the effective mass matrix of the neutrino-antineutrino systems. Our findings indicate that the gravitational field, expressed in BL coordinates, significantly influences the transition probabilities in two-flavor oscillations of the neutrino-antineutrino system. We then apply the expression for transition probabilities between states to analyze the Bures angle, which quantifies the closeness between the initial and final states of the time-evolved flavor state. We use this concept to probe the QSL for the time evolution of the initial flavor neutrino state., Comment: 16 pages, 12 figures, Accepted for publication in the Proceedings of the 17th Marcel Grossman Meeting (MG17), Pescara, Italy, 7-12 July 2024. Based on the talk presented in the parallel session "Unveiling neutrino secrets through cosmology: current status and future developments (NU2)"

The quantum speed limits (QSLs) determine the minimal amount of time required for a quantum system to evolve from an initial to a final state. We investigate QSLs for the unitary evolution of the neutrino-antineutrino system in the presence of a gravitational field. It is known that the transition probabilities between neutrino and antineutrino in the framework of one and two flavors depend on the strength of the gravitational field. The behavior of the QSL time in the two-flavor system indicates fast flavor transitions as the gravitational field strength increases. Subsequently, we observe quick suppression of entanglement by exploring the speed limit for entanglement entropy of two-flavor oscillations in the neutrino-antineutrino system in the proximity of a spinning primordial black hole., Comment: 12 pages, 5 figures, To be published in Astrophysics and Space Science Proceedings, titled "The Relativistic Universe: From Classical to Quantum, Proceedings of the International Symposium on Recent Developments in Relativistic Astrophysics", Gangtok, December 11-13, 2023: to felicitate Prof. Banibrata Mukhopadhyay on his 50th Birth Anniversary", Editors: S Ghosh & A R Rao, Springer Nature

As the outputs of generative AI (GenAI) techniques improve in quality, it becomes increasingly challenging to distinguish them from human-created content. Watermarking schemes are a promising approach to address the problem of distinguishing between AI and human-generated content. These schemes embed hidden signals within AI-generated content to enable reliable detection. While watermarking is not a silver bullet for addressing all risks associated with GenAI, it can play a crucial role in enhancing AI safety and trustworthiness by combating misinformation and deception. This paper presents a comprehensive overview of watermarking techniques for GenAI, beginning with the need for watermarking from historical and regulatory perspectives. We formalize the definitions and desired properties of watermarking schemes and examine the key objectives and threat models for existing approaches. Practical evaluation strategies are also explored, providing insights into the development of robust watermarking techniques capable of resisting various attacks. Additionally, we review recent representative works, highlight open challenges, and discuss potential directions for this emerging field. By offering a thorough understanding of watermarking in GenAI, this work aims to guide researchers in advancing watermarking methods and applications, and support policymakers in addressing the broader implications of GenAI.

Active Learning aims to optimize performance while minimizing annotation costs by selecting the most informative samples from an unlabelled pool. Traditional uncertainty sampling often leads to sampling bias by choosing similar uncertain samples. We propose an active learning method that utilizes fixed equiangular hyperspherical points as class prototypes, ensuring consistent inter-class separation and robust feature representations. Our approach introduces Maximally Separated Active Learning (MSAL) for uncertainty sampling and a combined strategy (MSAL-D) for incorporating diversity. This method eliminates the need for costly clustering steps, while maintaining diversity through hyperspherical uniformity. We demonstrate strong performance over existing active learning techniques across five benchmark datasets, highlighting the method's effectiveness and integration ease. The code is available on GitHub., Comment: ECCV 2024 Beyond Euclidean Workshop (proceedings)

This paper demonstrates that hedge funds tend to design their activist campaigns to align with the preferences and ideologies of institutions holding large stakes in the target company. I estimate these preferences by analyzing the institutions' previous proxy voting behavior. The results reveal that activists benefit from this approach. Campaigns with a stronger positive correlation between the preferences of larger institutions and activist communications attract more shareholder attention, receive more votes, and are more likely to succeed., Comment: 28 pages, 5 figures, 10 tables

We study scattering in Ising Field Theory (IFT) using matrix product states and the time-dependent variational principle. IFT is a one-parameter family of strongly coupled non-integrable quantum field theories in 1+1 dimensions, interpolating between massive free fermion theory and Zamolodchikov's integrable massive $E_8$ theory. Particles in IFT may scatter either elastically or inelastically. In the post-collision wavefunction, particle tracks from all final-state channels occur in superposition; processes of interest can be isolated by projecting the wavefunction onto definite particle sectors, or by evaluating energy density correlation functions. Using numerical simulations we determine the time delay of elastic scattering and the probability of inelastic particle production as a function of collision energy. We also study the mass and width of the lightest resonance near the $E_8$ point in detail. Close to both the free fermion and $E_8$ theories, our results for both elastic and inelastic scattering are in good agreement with expectations from form-factor perturbation theory. Using numerical computations to go beyond the regime accessible by perturbation theory, we find that the high energy behavior of the two-to-two particle scattering probability in IFT is consistent with a conjecture of Zamolodchikov. Our results demonstrate the efficacy of tensor-network methods for simulating the real-time dynamics of strongly coupled quantum field theories in 1+1 dimensions., Comment: 16 + 12 pages, many spacetime pictures of scattering processes

Efficient path optimization for drones in search and rescue operations faces challenges, including limited visibility, time constraints, and complex information gathering in urban environments. We present a comprehensive approach to optimize UAV-based search and rescue operations in neighborhood areas, utilizing both a 3D AirSim-ROS2 simulator and a 2D simulator. The path planning problem is formulated as a partially observable Markov decision process (POMDP), and we propose a novel ``Shrinking POMCP'' approach to address time constraints. In the AirSim environment, we integrate our approach with a probabilistic world model for belief maintenance and a neurosymbolic navigator for obstacle avoidance. The 2D simulator employs surrogate ROS2 nodes with equivalent functionality. We compare trajectories generated by different approaches in the 2D simulator and evaluate performance across various belief types in the 3D AirSim-ROS simulator. Experimental results from both simulators demonstrate that our proposed shrinking POMCP solution achieves significant improvements in search times compared to alternative methods, showcasing its potential for enhancing the efficiency of UAV-assisted search and rescue operations., Comment: Accepted to the The 3rd International Conference on Assured Autonomy

The question of thermalization of a closed quantum system is of central interest in non-equilibrium quantum many-body physics. Here we present one such study analyzing the dynamics of a closed coupled Majorana SYK system. We have a large-$q$ SYK model prepared initially at equilibrium quenched by introducing a random hopping term, thus leading to non-equilibrium dynamics. We find that the final stationary state reaches thermal equilibrium with respect to the Green's functions and energy. Accordingly, the final state is characterized by calculating its final temperature and the thermalization rate. We provide a detailed review of analytical methods and derive the required Kadanoff-Baym equations, which are then solved using the algorithm developed in this work. Our results display rich thermalization dynamics in a closed quantum system in the thermodynamic limit., Comment: 31 pages, 15 figures

The rapid adoption of e-commerce has transformed how students access goods and resources. However, existing platforms often fail to address the specific needs of campus communities, where students face challenges such as financial constraints, lack of access to affordable goods, and inefficient resource circulation. This research proposes ShareSpace, a localized web application designed specifically for college students to facilitate the buying, and selling of mainly second-hand goods. By addressing imbalances like surplus items left behind by seniors and shortages experienced by juniors, ShareSpace promotes sustainability and affordability within the campus ecosystem. Leveraging modern technologies such as Node.js, React.js, and MongoDB, the project demonstrates the feasibility of creating a student-centric e-commerce solution. The study highlights how ShareSpace solves the challenges of economical pricing and content moderation using proposed solutions. This study also explores the limitations of existing solutions and evaluates the potential of ShareSpace to encourage sustainable consumption and resourcefulness among students.

Electron-phonon (E-p) coupling incorporated density functional theory (DFT) based investigation of structural, electronic and vibrational properties of bulk MgSe and MgTe is presented. Electron-phonon coupling is incorporated to understand its effect on charge carrier dynamics. It is observed that the MgTe possesses room temperature hole and electron mobility of 18.7 cm2/Vs and 335 cm2/Vs, respectively; in contrast to this, the bulk MgSe follows reverse trend in temperature dependent carrier mobilities owing to its different scattering rate and electron-phonon coupling profiles. The key feature of the study was to showcase the importance of electron-phonon coupling in determining the carrier mobility and the relative dynamics in the material. Further, the incorporation of e-p coupling softens the electronic and phonon dispersions which is subjected to the inclusion of the interaction between the electrons and the phonons of the systems. The overall results indicate that the incorporation of the electron-phonon coupling is crucial in determining the carrier dynamics of the system which is in excellent agreement with the experimental findings. Both materials possess moderate magnitudes of mobilities that are subjected to the modification by means of changing the dimensional confinement and/or chemical composition that can strongly influence the carrier dynamics by means of altered edge states and coupling parameters., Comment: 19 pages, 3 Figures

Over an extensive duration, administrators and clinicians have endeavoured to predict Emergency Department (ED) visits with precision, aiming to optimise resource distribution. Despite the proliferation of diverse AI-driven models tailored for precise prognostication, this task persists as a formidable challenge, besieged by constraints such as restrained generalisability, susceptibility to overfitting and underfitting, scalability issues, and complex fine-tuning hyper-parameters. In this study, we introduce a novel Meta-learning Gradient Booster (Meta-ED) approach for precisely forecasting daily ED visits and leveraging a comprehensive dataset of exogenous variables, including socio-demographic characteristics, healthcare service use, chronic diseases, diagnosis, and climate parameters spanning 23 years from Canberra Hospital in ACT, Australia. The proposed Meta-ED consists of four foundational learners-Catboost, Random Forest, Extra Tree, and lightGBoost-alongside a dependable top-level learner, Multi-Layer Perceptron (MLP), by combining the unique capabilities of varied base models (sub-learners). Our study assesses the efficacy of the Meta-ED model through an extensive comparative analysis involving 23 models. The evaluation outcomes reveal a notable superiority of Meta-ED over the other models in accuracy at 85.7% (95% CI ;85.4%, 86.0%) and across a spectrum of 10 evaluation metrics. Notably, when compared with prominent techniques, XGBoost, Random Forest (RF), AdaBoost, LightGBoost, and Extra Tree (ExT), Meta-ED showcases substantial accuracy enhancements of 58.6%, 106.3%, 22.3%, 7.0%, and 15.7%, respectively. Furthermore, incorporating weather-related features demonstrates a 3.25% improvement in the prediction accuracy of visitors' numbers. The encouraging outcomes of our study underscore Meta-ED as a foundation model for the precise prediction of daily ED visitors.

Exascale computers offer transformative capabilities to combine data-driven and learning-based approaches with traditional simulation applications to accelerate scientific discovery and insight. However, these software combinations and integrations are difficult to achieve due to the challenges of coordinating and deploying heterogeneous software components on diverse and massive platforms. We present the ExaWorks project, which addresses many of these challenges. We developed a workflow Software Development Toolkit (SDK), a curated collection of workflow technologies that can be composed and interoperated through a common interface, engineered following current best practices, and specifically designed to work on HPC platforms. ExaWorks also developed PSI/J, a job management abstraction API, to simplify the construction of portable software components and applications that can be used over various HPC schedulers. The PSI/J API is a minimal interface for submitting and monitoring jobs and their execution state across multiple and commonly used HPC schedulers. We also describe several leading and innovative workflow examples of ExaWorks tools used on DOE leadership platforms. Furthermore, we discuss how our project is working with the workflow community, large computing facilities, and HPC platform vendors to address the requirements of workflows sustainably at the exascale.

We introduce LLMStinger, a novel approach that leverages Large Language Models (LLMs) to automatically generate adversarial suffixes for jailbreak attacks. Unlike traditional methods, which require complex prompt engineering or white-box access, LLMStinger uses a reinforcement learning (RL) loop to fine-tune an attacker LLM, generating new suffixes based on existing attacks for harmful questions from the HarmBench benchmark. Our method significantly outperforms existing red-teaming approaches (we compared against 15 of the latest methods), achieving a +57.2% improvement in Attack Success Rate (ASR) on LLaMA2-7B-chat and a +50.3% ASR increase on Claude 2, both models known for their extensive safety measures. Additionally, we achieved a 94.97% ASR on GPT-3.5 and 99.4% on Gemma-2B-it, demonstrating the robustness and adaptability of LLMStinger across open and closed-source models., Comment: Accepted at AAAI 2025

White matter tract segmentation is crucial for studying brain structural connectivity and neurosurgical planning. However, segmentation remains challenging due to issues like class imbalance between major and minor tracts, structural similarity, subject variability, symmetric streamlines between hemispheres etc. To address these challenges, we propose TractoEmbed, a modular multi-level embedding framework, that encodes localized representations through learning tasks in respective encoders. In this paper, TractoEmbed introduces a novel hierarchical streamline data representation that captures maximum spatial information at each level i.e. individual streamlines, clusters, and patches. Experiments show that TractoEmbed outperforms state-of-the-art methods in white matter tract segmentation across different datasets, and spanning various age groups. The modular framework directly allows the integration of additional embeddings in future works., Comment: Accepted at 27th International Conference on Pattern Recognition (ICPR), 2024 15 pages, 2 figures

We present analytical and simulation study of twisted terahertz (THz) radiation generation via propagation of a circularly polarized Laguerre Gaussian (LG) laser pulse in homogeneous plasma embedded in an axial magnetic field. Analytical formulation is based on perturbation technique and quasistatic approximation. Longitudinal and transverse wakefields generated via laser plasma interactions are evaluated using Lorentz force and Maxwells equations in the mildly nonlinear regime. It is observed that two linearly polarized twisted terahertz (THz) radiation beams are generated in mutually perpendicular planes. Superposition of the two beams result in a single linearly polarized twisted THz radiation beam with modified amplitude and polarization direction. Three dimensional (3D) particle in cell (PIC) simulations are performed for this configuration using FBPIC code. Graphical comparison of amplitude of the resultant THz beam obtained via analytical and simulation studies is presented.

Fiber tractography is a cornerstone of neuroimaging, enabling the detailed mapping of the brain's white matter pathways through diffusion MRI. This is crucial for understanding brain connectivity and function, making it a valuable tool in neurological applications. Despite its importance, tractography faces challenges due to its complexity and susceptibility to false positives, misrepresenting vital pathways. To address these issues, recent strategies have shifted towards deep learning, utilizing supervised learning, which depends on precise ground truth, or reinforcement learning, which operates without it. In this work, we propose Tract-RLFormer, a network utilizing both supervised and reinforcement learning, in a two-stage policy refinement process that markedly improves the accuracy and generalizability across various data-sets. By employing a tract-specific approach, our network directly delineates the tracts of interest, bypassing the traditional segmentation process. Through rigorous validation on datasets such as TractoInferno, HCP, and ISMRM-2015, our methodology demonstrates a leap forward in tractography, showcasing its ability to accurately map the brain's white matter tracts., Comment: Accepted at 27th International Conference on Pattern Recognition (ICPR), 2024

Large-scale foundation models like CLIP have shown strong zero-shot generalization but struggle with domain shifts, limiting their adaptability. In our work, we introduce \textsc{StyLIP}, a novel domain-agnostic prompt learning strategy for Domain Generalization (DG). StyLIP disentangles visual style and content in CLIP`s vision encoder by using style projectors to learn domain-specific prompt tokens and combining them with content features. Trained contrastively, this approach enables seamless adaptation across domains, outperforming state-of-the-art methods on multiple DG benchmarks. Additionally, we propose AD-CLIP for unsupervised domain adaptation (DA), leveraging CLIP`s frozen vision backbone to learn domain-invariant prompts through image style and content features. By aligning domains in embedding space with entropy minimization, AD-CLIP effectively handles domain shifts, even when only target domain samples are available. Lastly, we outline future work on class discovery using prompt learning for semantic segmentation in remote sensing, focusing on identifying novel or rare classes in unstructured environments. This paves the way for more adaptive and generalizable models in complex, real-world scenarios., Comment: ICVGIP 2024, Young Faculty Symposium

We present a comprehensive multi-epoch photometric and spectroscopic study of SN 2024bch, a nearby (19.9 Mpc) Type II supernova (SN) with prominent early high ionization emission lines. Optical spectra from 2.9 days after the estimated explosion reveal narrow lines of H I, He II, C IV, and N IV that disappear by day 6. High cadence photometry from the ground and TESS show that the SN brightened quickly and reached a peak M$_V \sim$ $-$17.8 mag within a week of explosion, and late-time photometry suggests a $^{56}$Ni mass of 0.050 M$_{\odot}$. High-resolution spectra from day 8 and 43 trace the unshocked circumstellar medium (CSM) and indicate a wind velocity of 30--40 km s$^{-1}$, a value consistent with a red supergiant (RSG) progenitor. Comparisons between models and the early spectra suggest a pre-SN mass-loss rate of $\dot{M} \sim 10^{-3}-10^{-2}\ M_\odot\ \mathrm{yr}^{-1}$, which is too high to be explained by quiescent mass loss from RSGs, but is consistent with some recent measurements of similar SNe. Persistent blueshifted H I and [O I] emission lines seen in the optical and NIR spectra could be produced by asymmetries in the SN ejecta, while the multi-component H$\alpha$ may indicate continued interaction with an asymmetric CSM well into the nebular phase. SN 2024bch provides another clue to the complex environments and mass-loss histories around massive stars., Comment: Submitted to ApJ

We present supernova (SN) 2023ufx, a unique Type IIP SN with the shortest known plateau duration ($t_\mathrm{PT}$ $\sim$47 days), a luminous V-band peak ($M_{V}$ = $-$18.42 $\pm$ 0.08 mag), and a rapid early decline rate ($s1$ = 3.47 $\pm$ 0.09 mag (50 days)$^{-1}$). By comparing observed photometry to a hydrodynamic MESA+STELLA model grid, we constrain the progenitor to be a massive red supergiant with M$_\mathrm{ZAMS}$ $\simeq$19 - 25 M$_{\odot}$. Independent comparisons with nebular spectral models also suggest an initial He-core mass of $\sim$6 M$_{\odot}$, and thus a massive progenitor. For a Type IIP, SN 2023ufx produced an unusually high amount of nickel ($^{56}$Ni) $\sim$0.14 $\pm$ 0.02 M$_{\odot}$, during the explosion. We find that the short plateau duration in SN 2023ufx can be explained with the presence of a small hydrogen envelope (M$_\mathrm{H_\mathrm{env}}$ $\simeq$1.2 M$_{\odot}$), suggesting partial stripping of the progenitor. About $\simeq$0.09 M$_{\odot}$ of CSM through mass loss from late-time stellar evolution of the progenitor is needed to fit the early time ($\lesssim$10 days) pseudo-bolometric light curve. Nebular line diagnostics of broad and multi-peak components of [O I] $\lambda\lambda$6300, 6364, H$\alpha$, and [Ca II] $\lambda \lambda$7291, 7323 suggest that the explosion of SN 2023ufx could be inherently asymmetric, preferentially ejecting material along our line-of-sight., Comment: Submitted to ApJ, 30 pages, 19 figures

In this paper, we present a dynamic semantic clustering approach inspired by the Chinese Restaurant Process, aimed at addressing uncertainty in the inference of Large Language Models (LLMs). We quantify uncertainty of an LLM on a given query by calculating entropy of the generated semantic clusters. Further, we propose leveraging the (negative) likelihood of these clusters as the (non)conformity score within Conformal Prediction framework, allowing the model to predict a set of responses instead of a single output, thereby accounting for uncertainty in its predictions. We demonstrate the effectiveness of our uncertainty quantification (UQ) technique on two well known question answering benchmarks, COQA and TriviaQA, utilizing two LLMs, Llama2 and Mistral. Our approach achieves SOTA performance in UQ, as assessed by metrics such as AUROC, AUARC, and AURAC. The proposed conformal predictor is also shown to produce smaller prediction sets while maintaining the same probabilistic guarantee of including the correct response, in comparison to existing SOTA conformal prediction baseline.

Generalized Category Discovery (GCD) aims to cluster unlabeled images into known and novel categories using labeled images from known classes. To address the challenge of transferring features from known to unknown classes while mitigating model bias, we introduce GraphVL, a novel approach for vision-language modeling in GCD, leveraging CLIP. Our method integrates a graph convolutional network (GCN) with CLIP's text encoder to preserve class neighborhood structure. We also employ a lightweight visual projector for image data, ensuring discriminative features through margin-based contrastive losses for image-text mapping. This neighborhood preservation criterion effectively regulates the semantic space, making it less sensitive to known classes. Additionally, we learn textual prompts from known classes and align them to create a more contextually meaningful semantic feature space for the GCN layer using a contextual similarity loss. Finally, we represent unlabeled samples based on their semantic distance to class prompts from the GCN, enabling semi-supervised clustering for class discovery and minimizing errors. Our experiments on seven benchmark datasets consistently demonstrate the superiority of GraphVL when integrated with the CLIP backbone., Comment: Accepted in ACM ICVGIP 2024

We introduce a new identification strategy for uncertainty shocks to explain macroeconomic volatility in financial markets. The Chicago Board Options Exchange Volatility Index (VIX) measures market expectations of future volatility, but traditional methods based on second-moment shocks and time-varying volatility of the VIX often fail to capture the non-Gaussian, heavy-tailed nature of asset returns. To address this, we construct a revised VIX by fitting a double-subordinated Normal Inverse Gaussian Levy process to S&P 500 option prices, providing a more comprehensive measure of volatility that reflects the extreme movements and heavy tails observed in financial data. Using an axiomatic approach, we introduce a general family of risk-reward ratios, computed with our revised VIX and fitted over a fractional time series to more accurately identify uncertainty shocks in financial markets.

Transformers are set to become ubiquitous with applications ranging from chatbots and educational assistants to visual recognition and remote sensing. However, their increasing computational and memory demands is resulting in growing energy consumption. Building models with fast and energy-efficient inference is imperative to enable a variety of transformer-based applications. Look Up Table (LUT) based Weightless Neural Networks are faster than the conventional neural networks as their inference only involves a few lookup operations. Recently, an approach for learning LUT networks directly via an Extended Finite Difference method was proposed. We build on this idea, extending it for performing the functions of the Multi Layer Perceptron (MLP) layers in transformer models and integrating them with transformers to propose Quasi Weightless Transformers (QuWeiT). This allows for a computational and energy-efficient inference solution for transformer-based models. On I-ViT-T, we achieve a comparable accuracy of 95.64% on CIFAR-10 dataset while replacing approximately 55% of all the multiplications in the entire model and achieving a 2.2x energy efficiency. We also observe similar savings on experiments with the nanoGPT framework.

Additive Manufacturing (AM) is transforming the manufacturing sector by enabling efficient production of intricately designed products and small-batch components. However, metal parts produced via AM can include flaws that cause inferior mechanical properties, including reduced fatigue response, yield strength, and fracture toughness. To address this issue, we leverage convolutional neural networks (CNN) to analyze thermal images of printed layers, automatically identifying anomalies that impact these properties. We also investigate various synthetic data generation techniques to address limited and imbalanced AM training data. Our models' defect detection capabilities were assessed using images of Nickel alloy 718 layers produced on a laser powder bed fusion AM machine and synthetic datasets with and without added noise. Our results show significant accuracy improvements with synthetic data, emphasizing the importance of expanding training sets for reliable defect detection. Specifically, Generative Adversarial Networks (GAN)-generated datasets streamlined data preparation by eliminating human intervention while maintaining high performance, thereby enhancing defect detection capabilities. Additionally, our denoising approach effectively improves image quality, ensuring reliable defect detection. Finally, our work integrates these models in the CLoud ADditive MAnufacturing (CLADMA) module, a user-friendly interface, to enhance their accessibility and practicality for AM applications. This integration supports broader adoption and practical implementation of advanced defect detection in AM processes., Comment: 29 pages

Ultrasonic Additive Manufacturing (UAM) employs ultrasonic welding to bond similar or dissimilar metal foils to a substrate, resulting in solid, consolidated metal components. However, certain processing conditions can lead to inter-layer defects, affecting the final product's quality. This study develops a method to monitor in-process quality using deep learning-based convolutional neural networks (CNNs). The CNN models were evaluated on their ability to classify samples with and without embedded thermocouples across five power levels (300W, 600W, 900W, 1200W, 1500W) using thermal images with supervised labeling. Four distinct CNN classification models were created for different scenarios including without (baseline) and with thermocouples, only without thermocouples across power levels, only with thermocouples across power levels, and combined without and with thermocouples across power levels. The models achieved 98.29% accuracy on combined baseline and thermocouple images, 97.10% for baseline images across power levels, 97.43% for thermocouple images, and 97.27% for both types across power levels. The high accuracy, above 97%, demonstrates the system's effectiveness in identifying and classifying conditions within the UAM process, providing a reliable tool for quality assurance and process control in manufacturing environments.

We report the discovery of a peculiar X-ray transient, EP240408a, by Einstein Probe (EP) and follow-up studies made with EP, Swift, NICER, GROND, ATCA and other ground-based multi-wavelength telescopes. The new transient was first detected with Wide-field X-ray Telescope (WXT) on board EP on April 8th, 2024, manifested in an intense yet brief X-ray flare lasting for 12 seconds. The flare reached a peak flux of 3.9x10^(-9) erg/cm2/s in 0.5-4 keV, about 300 times brighter than the underlying X-ray emission detected throughout the observation. Rapid and more precise follow-up observations by EP/FXT, Swift and NICER confirmed the finding of this new transient. Its X-ray spectrum is non-thermal in 0.5-10 keV, with a power-law photon index varying within 1.8-2.5. The X-ray light curve shows a plateau lasting for about 4 days, followed by a steep decay till becoming undetectable about 10 days after the initial detection. Based on its temporal property and constraints from previous EP observations, an unusual timescale in the range of 7-23 days is found for EP240408a, which is intermediate between the commonly found fast and long-term transients. No counterparts have been found in optical and near-infrared, with the earliest observation at 17 hours after the initial X-ray detection, suggestive of intrinsically weak emission in these bands. We demonstrate that the remarkable properties of EP240408a are inconsistent with any of the transient types known so far, by comparison with, in particular, jetted tidal disruption events, gamma-ray bursts, X-ray binaries and fast blue optical transients. The nature of EP240408a thus remains an enigma. We suggest that EP240408a may represent a new type of transients with intermediate timescales of the order of about 10 days. The detection and follow-ups of more of such objects are essential for revealing their origin., Comment: 25 pages, 11 figures

Training data compositions for Large Language Models (LLMs) can significantly affect their downstream performance. However, a thorough data ablation study exploring large sets of candidate data mixtures is typically prohibitively expensive since the full effect is seen only after training the models; this can lead practitioners to settle for sub-optimal data mixtures. We propose an efficient method for approximating data ablations which trains individual models on subsets of a training corpus and reuses them across evaluations of combinations of subsets. In continued pre-training experiments, we find that, given an arbitrary evaluation set, the perplexity score of a single model trained on a candidate set of data is strongly correlated with perplexity scores of parameter averages of models trained on distinct partitions of that data. From this finding, we posit that researchers and practitioners can conduct inexpensive simulations of data ablations by maintaining a pool of models that were each trained on partitions of a large training corpus, and assessing candidate data mixtures by evaluating parameter averages of combinations of these models. This approach allows for substantial improvements in amortized training efficiency -- scaling only linearly with respect to new data -- by enabling reuse of previous training computation, opening new avenues for improving model performance through rigorous, incremental data assessment and mixing., Comment: EMNLP 2024. 17 pages

The Workflows Community Summit gathered 111 participants from 18 countries to discuss emerging trends and challenges in scientific workflows, focusing on six key areas: time-sensitive workflows, AI-HPC convergence, multi-facility workflows, heterogeneous HPC environments, user experience, and FAIR computational workflows. The integration of AI and exascale computing has revolutionized scientific workflows, enabling higher-fidelity models and complex, time-sensitive processes, while introducing challenges in managing heterogeneous environments and multi-facility data dependencies. The rise of large language models is driving computational demands to zettaflop scales, necessitating modular, adaptable systems and cloud-service models to optimize resource utilization and ensure reproducibility. Multi-facility workflows present challenges in data movement, curation, and overcoming institutional silos, while diverse hardware architectures require integrating workflow considerations into early system design and developing standardized resource management tools. The summit emphasized improving user experience in workflow systems and ensuring FAIR workflows to enhance collaboration and accelerate scientific discovery. Key recommendations include developing standardized metrics for time-sensitive workflows, creating frameworks for cloud-HPC integration, implementing distributed-by-design workflow modeling, establishing multi-facility authentication protocols, and accelerating AI integration in HPC workflow management. The summit also called for comprehensive workflow benchmarks, workflow-specific UX principles, and a FAIR workflow maturity model, highlighting the need for continued collaboration in addressing the complex challenges posed by the convergence of AI, HPC, and multi-facility research environments.

We present a quantum computational framework for SU(2) lattice gauge theory, leveraging continuous variables instead of discrete qubits to represent the infinite-dimensional Hilbert space of the gauge fields. We consider a ladder as well as a two-dimensional grid of plaquettes, detailing the use of gauge fixing to reduce the degrees of freedom and simplify the Hamiltonian. We demonstrate how the system dynamics, ground states, and energy gaps can be computed using the continuous-variable approach to quantum computing. Our results indicate that it is feasible to study non-Abelian gauge theories with continuous variables, providing new avenues for understanding the real-time dynamics of quantum field theories., Comment: 26 pages

In this article, we study a two-dimensional singularly perturbed parabolic equation of the convection-diffusion type, characterized by discontinuities in the source term and convection coefficient at a specific point in the domain. These discontinuities lead to the development of interior layers. To address these layers and ensure uniform convergence, we propose a hybrid monotone difference scheme that combines the central difference and midpoint upwind schemes for spatial discretization, applied on a piecewise-uniform Shishkin mesh. For temporal discretization, we employ the Crank-Nicolson method on a uniform mesh. The resulting scheme is proven to be uniformly convergent, order achieving almost two in space and two in time. Numerical experiments validate the theoretical error estimates, demonstrating superior accuracy and convergence when compared to existing methods., Comment: arXiv admin note: substantial text overlap with arXiv:2409.00354

Model-based planners and controllers are commonly used to solve complex manipulation problems as they can efficiently optimize diverse objectives and generalize to long horizon tasks. However, they are limited by the fidelity of their model which oftentimes leads to failures during deployment. To enable a robot to recover from such failures, we propose to use hierarchical reinforcement learning to learn a separate recovery policy. The recovery policy is triggered when a failure is detected based on sensory observations and seeks to take the robot to a state from which it can complete the task using the nominal model-based controllers. Our approach, called RecoveryChaining, uses a hybrid action space, where the model-based controllers are provided as additional \emph{nominal} options which allows the recovery policy to decide how to recover, when to switch to a nominal controller and which controller to switch to even with \emph{sparse rewards}. We evaluate our approach in three multi-step manipulation tasks with sparse rewards, where it learns significantly more robust recovery policies than those learned by baselines. Finally, we successfully transfer recovery policies learned in simulation to a physical robot to demonstrate the feasibility of sim-to-real transfer with our method., Comment: 8 pages, 9 figures

The pervasiveness of proprietary language models has raised privacy concerns for users' sensitive data, emphasizing the need for private inference (PI), where inference is performed directly on encrypted inputs. However, current PI methods face prohibitively higher communication and latency overheads, primarily due to nonlinear operations. In this paper, we present a comprehensive analysis to understand the role of nonlinearities in transformer-based decoder-only language models. We introduce AERO, a four-step architectural optimization framework that refines the existing LLM architecture for efficient PI by systematically removing nonlinearities such as LayerNorm and GELU and reducing FLOPs counts. For the first time, we propose a Softmax-only architecture with significantly fewer FLOPs tailored for efficient PI. Furthermore, we devise a novel entropy regularization technique to improve the performance of Softmax-only models. AERO achieves up to 4.23$\times$ communication and 1.94$\times$ latency reduction. We validate the effectiveness of AERO by benchmarking it against the state-of-the-art., Comment: 40 pages, 21 figures, and 9 tables

Recent works have shown that Federated Learning (FL) is vulnerable to backdoor attacks. Existing defenses cluster submitted updates from clients and select the best cluster for aggregation. However, they often rely on unrealistic assumptions regarding client submissions and sampled clients population while choosing the best cluster. We show that in realistic FL settings, state-of-the-art (SOTA) defenses struggle to perform well against backdoor attacks in FL. To address this, we highlight that backdoored submissions are adversarially biased and overconfident compared to clean submissions. We, therefore, propose an Adversarially Guided Stateful Defense (AGSD) against backdoor attacks on Deep Neural Networks (DNNs) in FL scenarios. AGSD employs adversarial perturbations to a small held-out dataset to compute a novel metric, called the trust index, that guides the cluster selection without relying on any unrealistic assumptions regarding client submissions. Moreover, AGSD maintains a trust state history of each client that adaptively penalizes backdoored clients and rewards clean clients. In realistic FL settings, where SOTA defenses mostly fail to resist attacks, AGSD mostly outperforms all SOTA defenses with minimal drop in clean accuracy (5% in the worst-case compared to best accuracy) even when (a) given a very small held-out dataset -- typically AGSD assumes 50 samples (<= 0.1% of the training data) and (b) no heldout dataset is available, and out-of-distribution data is used instead. For reproducibility, our code will be openly available at: https://github.com/hassanalikhatim/AGSD., Comment: 16 pages, Accepted at ACSAC 2024