Search

We give an exact application of a recently developed, operator-based theory of wave propagation in dispersive, time-varying media. Using this theory we find that the usual symmetry of complex conjugation plus changing the sign of the frequency, required for real valued fields, implies that the allowed propagation constants in the medium are either real valued or come in conjugate pairs. The real valued wave numbers are only present in time-varying media, implying that time variation leads to modes that are free from dissipation, even in a lossy medium. Moreover, these symmetry-unbroken waves lack a defined propagation direction. This can lead to a divergent transmission coefficient when waves are incident onto a finite, time-varying slab. The techniques used in this work present a route towards further analytic applications of this operator formalism., Comment: 23 page, 6 figures

Emerging Large Language Model (LLM) applications require long input prompts to perform complex downstream tasks like document analysis and code generation. For these long context length applications, the length of the input prompt poses a significant challenge in terms of inference efficiency since the inference costs increase linearly with sequence length. However, for many of these applications, much of the context in the prompt is fixed across different user inputs, thereby providing the opportunity to perform offline optimizations to process user inputs quickly, as they are received. In this work, we propose Squeezed Attention as a mechanism to accelerate LLM applications where a large portion of the input prompt is fixed. We first leverage K-means clustering offline to group the keys for the fixed context based on semantic similarity and represent each cluster with a single centroid value. During inference, we compare query tokens from the user input with the centroids to predict which of the keys from the fixed context are semantically relevant and need to be loaded during inference. We then compute exact attention using only these important keys from the fixed context, thereby reducing bandwidth and computational costs. We also extend our method to use a hierarchical centroid lookup to identify important keys, which can reduce the complexity of attention from linear to logarithmic with respect to the context length. We implement optimized Triton kernels for centroid comparison and sparse FlashAttention with important keys, achieving more than 4x speedups during both the prefill and generation phases for long-context inference. Furthermore, we have extensively evaluated our method on various long-context benchmarks including LongBench, where it achieves a 3x reduction in KV cache budget without accuracy loss and up to an 8x reduction with <0.5 point accuracy gap for various models.

Active metamaterials are engineered structures that possess novel properties that can be changed after the point of manufacture. Their novel properties arise predominantly from their physical structure, as opposed to their chemical composition and can be changed through means such as direct energy addition into wave paths, or physically changing/morphing the structure in response to both a user or environmental input. Active metamaterials are currently of wide interest to the physics community and encompass a range of sub-domains in applied physics (e.g. photonic, microwave, acoustic, mechanical, etc.). They possess the potential to provide solutions that are more suitable to specific applications, or which allow novel properties to be produced which cannot be achieved with passive metamaterials, such as time-varying or gain enhancement effects. They have the potential to help solve some of the important current and future problems faced by the advancement of modern society, such as achieving net-zero, sustainability, healthcare and equality goals. Despite their huge potential, the added complexity of their design and operation, compared to passive metamaterials creates challenges to the advancement of the field, particularly beyond theoretical and lab-based experiments. This roadmap brings together experts in all types of active metamaterials and across a wide range of areas of applied physics. The objective is to provide an overview of the current state of the art and the associated current/future challenges, with the hope that the required advances identified create a roadmap for the future advancement and application of this field.

Recent large language models (LLMs) have enabled the development of advanced agentic systems that can integrate various tools and APIs to fulfill user queries through function calling. However, the deployment of these LLMs on the edge has not been explored since they typically require cloud-based infrastructure due to their substantial model size and computational demands. To this end, we present TinyAgent, an end-to-end framework for training and deploying task-specific small language model agents capable of function calling for driving agentic systems at the edge. We first show how to enable accurate function calling for open-source models via the LLMCompiler framework. We then systematically curate a high-quality dataset for function calling, which we use to fine-tune two small language models, TinyAgent-1.1B and 7B. For efficient inference, we introduce a novel tool retrieval method to reduce the input prompt length and utilize quantization to further accelerate the inference speed. As a driving application, we demonstrate a local Siri-like system for Apple's MacBook that can execute user commands through text or voice input. Our results show that our models can achieve, and even surpass, the function-calling capabilities of larger models like GPT-4-Turbo, while being fully deployed at the edge. We open-source our dataset, models, and installable package and provide a demo video for our MacBook assistant agent., Comment: EMNLP 2024 Demo

We show that a large family of tensorial metasurfaces can be found that perform an identical wave transformation, showing that even when the conditions of reciprocity and passivity are imposed, there still remain many solutions to the design problem. As an example, we explore the case of a metasurface that rotates a single input polarization, showing we can parameterize the set of equivalent reciprocal metasurfaces in terms of a single complex parameter. Through allowing dissipation and gain within the response, the surface can have many different functionalities in the orthogonal polarization, opening up a new route for the design of multiplexed metasurfaces.

In recent years, Large Language Models (LLM) have demonstrated impressive capabilities in the field of natural language processing (NLP). This paper explores the application of LLMs in negotiation transcript analysis by the Vanderbilt AI Negotiation Lab. Starting in September 2022, we applied multiple strategies using LLMs from zero shot learning to fine tuning models to in-context learning). The final strategy we developed is explained, along with how to access and use the model. This study provides a sense of both the opportunities and roadblocks for the implementation of LLMs in real life applications and offers a model for how LLMs can be applied to coding in other fields.

The interaction of light with superluminally moving matter entails unconventional phenomena, from Fresnel drag to Hawking radiation and to light amplification. While relativity makes these effects inaccessible using objects in motion, synthetic motion - enabled via space-time modulated internal degrees of freedom - is free from these constraints. Here we observe synthetic velocity of a reflectivity modulation travelling on an Indium-Tin-Oxide (ITO) interface, generated by ultrafast laser illumination at multiple positions and times. The interaction of the moving reflectivity modulation with a probe light beam acts as a non-separable spatio-temporal transformation that diffracts the light, changing its frequency and momentum content. The recorded frequency-momentum diffraction pattern is defined by the velocity of the diffracted probe wave relative to the modulation. Our experiments open a path towards mimicking relativistic mechanics and developing programmable spatio-temporal transformations of light., Comment: 9 pages, 3 figures

In models featuring extra spatial dimensions, particle collisions in the early universe can produce Kaluza-Klein gravitons. Such particles will later decay, potentially impacting the process of Big Bang nucleosynthesis. In this paper, we consider scenarios in which gravity is free to propagate throughout $n$ flat, compactified extra dimensions, while the fields of the Standard Model are confined to a 3+1 dimensional brane. We calculate the production and decay rates of the states that make up the Kaluza-Klein graviton tower and determine the evolution of their abundances in the early universe. We then go on to evaluate the impact of these decays on the resulting light element abundances. We identify significant regions of previously unexplored parameter space that are inconsistent with measurements of the primordial helium and deuterium abundances. In particular, we find that for the case of one extra dimension (two extra dimensions), the fundamental scale of gravity must be $M_{\star} > 2 \times 10^{13} \, {\rm GeV}$ ($M_{\star} > 10^{10} \, {\rm GeV}$) unless the temperature of the early universe was never greater than $T \sim 2 \, {\rm TeV}$ ($T \sim 1 \, {\rm GeV}$). For larger values of $n$, these constraints are less stringent. For the case of $n=6$, for example, our analysis excludes all values of $M_{\star}$ less than $\sim 10^{6} \, {\rm GeV}$, unless the temperature of the universe was never greater than $T \sim 3 \, {\rm TeV}$. The results presented here severely limit the possibility that black holes were efficiently produced through particle collisions in the early universe's thermal bath., Comment: 12 pages, 6 figures

Many lung diseases require detailed visualisation for accurate diagnosis and treatment. High-resolution computed tomography (CT) is the gold-standard technique for non-invasive lung disease detection, but it presents a risk to the patient through the relatively high ionising radiation dose required. Utilising the X-ray phase information may allow improvements in image resolution at equal or lower radiation levels than current clinical imaging. Propagation-based phase-contrast imaging requires minimal adaption of existing medical systems, and is well suited to lung imaging due to the strong phase gradients introduced by the lung-air material interfaces. Herein, propagation-based phase contrast CT is demonstrated for large animals, namely lambs, as a model for paediatric patients, using monochromatic radiation and a photon-counting detector at the Imaging and Medical Beamline of the Australian Synchrotron. Image quality, normalised against radiation dose, was optimised as a function of the beam energy and propagation distance, with the optimal conditions used to test the available image quality at very low radiation dose. The resulting CT images demonstrate superior resolution to existing high-resolution CT systems, pushing dose to the quantum limit to comply with current Australian guidelines for infant chest CT exposure of $<2.5\:\text{mSv}$ effective dose. Constituent raw projections are shown to have significant proportions of pixels with zero photon counts that would create severe information loss in conventional CT. Phase retrieval enabled clear visualisation of minor lung airways at doses up to 1,225$\pm$31\% times lower than conventional CT reconstruction, at a voxel size of just 75$\mathrm{\mu}$m.

Millicharged dark matter particles can be efficiently captured by the Sun, where they annihilate into tau leptons, leading to the production of high-energy neutrinos. In contrast to the Earth, the high temperature of the Sun suppresses the fraction of millicharged particles that are bound to nuclei, allowing for potentially high annihilation rates. We recast existing constraints from the IceCube Neutrino Observatory and use this information to place new limits on the fraction of the dark matter that is millicharged. This analysis excludes previously unexplored parameter space for masses of $m_\chi \sim (5-100) \ \text{GeV}$, charges of $q_\chi \sim 10^{-3}-10^{-2}$, and fractional abundances as small as $f_{_\text{DM}} \sim 10^{-5}$., Comment: 9 pages, 2 figures

Stochastic backgrounds of gravitational waves from primordial first-order phase transitions are a key probe of physics beyond the Standard Model. They represent one of the best prospects for observing or constraining new physics with the LISA gravitational wave observatory. However, the large foreground population of galactic binaries in the same frequency range represents a challenge, and will hinder the recovery of a stochastic background. To test the recoverability of a stochastic gravitational wave background, we use the LISA Simulation Suite to generate data incorporating both a stochastic background and an annually modulated foreground modelling the galactic binary population, and the Bayesian analysis code Cobaya to attempt to recover the model parameters. By applying the Deviance Information Criterion to compare models with and without a stochastic background we place bounds on the detectability of gravitational waves from first-order phase transitions. By further comparing models with and without the annual modulation, we show that exploiting the modulation improves the goodness-of-fit and gives a modest improvement to the bounds on detectable models., Comment: 29 pages, 9 figures

Pulsars convert a significant fraction of their total spin-down power into very high-energy electrons, leading to the formation of TeV halos. It is not yet known, however, whether these sources also efficiently accelerate electrons at lower energies and, if so, how those particles propagate through the surrounding environment. If pulsars produce $\sim 50-300 \, {\rm GeV}$ electrons, these particles would produce a spatially extended halo of synchrotron emission in the frequency range measured by Planck. Such emission could be used to constrain the low-energy diffusion coefficient in the regions surrounding these pulsars, as well as the spectrum and intensity of the electrons that are accelerated in this energy range. In this study, we attempt to use Planck data to constrain the nature of the Geminga pulsar's TeV halo. We find no conclusive evidence of this emission in Planck's frequency range, however, and calculate that the synchrotron flux from Geminga should be well below the total flux measured by Planck, even for models with favorable diffusion parameters or soft injection spectra. At this time, these measurements are not capable of significantly constraining the values of these parameters., Comment: 11 pages, 3 figures

Generative Artificial Intelligence (generative AI) poses both opportunities and risks for the integrity of research. Universities must guide researchers in using generative AI responsibly, and in navigating a complex regulatory landscape subject to rapid change. By drawing on the experiences of two Australian universities, we propose a framework to help institutions promote and facilitate the responsible use of generative AI. We provide guidance to help distil the diverse regulatory environment into a principles-based position statement. Further, we explain how a position statement can then serve as a foundation for initiatives in training, communications, infrastructure, and process change. Despite the growing body of literature about AI's impact on academic integrity for undergraduate students, there has been comparatively little attention on the impacts of generative AI for research integrity, and the vital role of institutions in helping to address those challenges. This paper underscores the urgency for research institutions to take action in this area and suggests a practical and adaptable framework for so doing.

The ability to recover MRI signal from noise is key to achieve fast acquisition, accurate quantification, and high image quality. Past work has shown convolutional neural networks can be used with abundant and paired low and high-SNR images for training. However, for applications where high-SNR data is difficult to produce at scale (e.g. with aggressive acceleration, high resolution, or low field strength), training a new denoising network using a large quantity of high-SNR images can be infeasible. In this study, we overcome this limitation by improving the generalization of denoising models, enabling application to many settings beyond what appears in the training data. Specifically, we a) develop a training scheme that uses complex MRIs reconstructed in the SNR units (i.e., the images have a fixed noise level, SNR unit training) and augments images with realistic noise based on coil g-factor, and b) develop a novel imaging transformer (imformer) to handle 2D, 2D+T, and 3D MRIs in one model architecture. Through empirical evaluation, we show this combination improves performance compared to CNN models and improves generalization, enabling a denoising model to be used across field-strengths, image contrasts, and anatomy.

The availability of unprecedented unsupervised training data, along with neural scaling laws, has resulted in an unprecedented surge in model size and compute requirements for serving/training LLMs. However, the main performance bottleneck is increasingly shifting to memory bandwidth. Over the past 20 years, peak server hardware FLOPS has been scaling at 3.0x/2yrs, outpacing the growth of DRAM and interconnect bandwidth, which have only scaled at 1.6 and 1.4 times every 2 years, respectively. This disparity has made memory, rather than compute, the primary bottleneck in AI applications, particularly in serving. Here, we analyze encoder and decoder Transformer models and show how memory bandwidth can become the dominant bottleneck for decoder models. We argue for a redesign in model architecture, training, and deployment strategies to overcome this memory limitation., Comment: Published in IEEE Micro Journal

It has been suggested that the Galactic Center Gamma-Ray Excess (GCE) could be produced by a large number of centrally-located millisecond pulsars. The fact that no such pulsar population has been detected implies that these sources must be very faint and very numerous. In this study, we use the contents of Fermi's recently released Third Pulsar Catalog (3PC) to measure the luminosity function of the millisecond pulsars in the Milky Way's Disk. We find that this source population exhibits a luminosity function with a mean gamma-ray luminosity of $\langle L_\gamma \rangle \sim 6 \times 10^{32}$ erg/s (integrated above 0.1 GeV). If the GCE were generated by millisecond pulsars with the same luminosity function, we find that $\sim$ 20 such sources from the Inner Galaxy population should have already been detected by Fermi and included in the 3PC. Given the lack of such observed sources, we exclude the hypothesis that the GCE is generated by pulsars with the same luminosity function as those in the Galactic Disk with a significance of $3.4\sigma$. We conclude that either less than 39% of the GCE is generated by pulsars, or that the millisecond pulsars in the Inner Galaxy are at least 5 times less luminous on average than those found in the Galactic Disk., Comment: 9 pages, 6 figures

To maintain full autonomy, autonomous robotic systems must have the ability to self-repair. Self-repairing via compensatory mechanisms appears in nature: for example, some fish can lose even 76% of their propulsive surface without loss of thrust by altering stroke mechanics. However, direct transference of these alterations from an organism to a robotic flapping propulsor may not be optimal due to irrelevant evolutionary pressures. We instead seek to determine what alterations to stroke mechanics are optimal for a damaged robotic system via artificial evolution. To determine whether natural and machine-learned optima differ, we employ a cyber-physical system using a Covariance Matrix Adaptation Evolutionary Strategy to seek the most efficient trajectory for a given force. We implement an online optimization with hardware-in-the-loop, performing experimental function evaluations with an actuated flexible flat plate. To recoup thrust production following partial amputation, the most efficient learned strategy was to increase amplitude, increase frequency, increase the amplitude of angle of attack, and phase shift the angle of attack by approximately 110 degrees. In fish, only an amplitude increase is reported by majority in the literature. To recoup side-force production, a more challenging optimization landscape is encountered. Nesting of optimal angle of attack traces is found in the resultant-based reference frame, but no clear trend in amplitude or frequency are exhibited -- in contrast to the increase in frequency reported in insect literature. These results suggest that how mechanical flapping propulsors most efficiently adjust to damage of a flapping propulsor may not align with natural swimmers and flyers.

LLMs are seeing growing use for applications which require large context windows, and with these large context windows KV cache activations surface as the dominant contributor to memory consumption during inference. Quantization is a promising approach for compressing KV cache activations; however, existing solutions fail to represent activations accurately in sub-4-bit precision. Our work, KVQuant, facilitates low precision KV cache quantization by incorporating several novel methods: (i) Per-Channel Key Quantization, where we adjust the dimension along which we quantize the Key activations to better match the distribution; (ii) Pre-RoPE Key Quantization, where we quantize Key activations before the rotary positional embedding to mitigate its impact on quantization; (iii) Non-Uniform KV Cache Quantization, where we derive per-layer sensitivity-weighted non-uniform datatypes that better represent the distributions; and (iv) Per-Vector Dense-and-Sparse Quantization, where we isolate outliers separately for each vector to minimize skews in quantization ranges. By applying our method to the LLaMA, Llama-2, Llama-3, and Mistral models, we achieve < 0.1 perplexity degradation with 3-bit quantization on both Wikitext-2 and C4, outperforming existing approaches. Our method enables serving LLaMA-7B with a context length of up to 1 million on a single A100-80GB GPU and up to 10 million on an 8-GPU system. We develop custom CUDA kernels for KVQuant, showing that we can achieve up to ~1.7x speedups, compared to baseline fp16 matrix-vector multiplications, for the LLaMA-7B model., Comment: NeurIPS 2024

Many applications must provide low-latency LLM service to users or risk unacceptable user experience. However, over-provisioning resources to serve fluctuating request patterns is often prohibitively expensive. In this work, we present a best-effort serving system that employs deep reinforcement learning to adjust service quality based on the task distribution and system load. Our best-effort system can maintain availability with over 10x higher client request rates, serves above 96% of peak performance 4.1x more often, and serves above 98% of peak performance 2.3x more often than static serving on unpredictable workloads. Our learned router is robust to shifts in both the arrival and task distribution. Compared to static serving, learned best-effort serving allows for cost-efficient serving through increased hardware utility. Additionally, we argue that learned best-effort LLM serving is applicable in wide variety of settings and provides application developers great flexibility to meet their specific needs., Comment: Es-FoMo @ ICML 2024

In this White Paper, we present recommendations for the scientific community and funding agencies to foster the infrastructure for a collaborative multi-messenger and time-domain astronomy (MMA/TDA) ecosystem. MMA/TDA is poised for breakthrough discoveries in the coming decade. In much the same way that expanding beyond the optical bandpass revealed entirely new and unexpected discoveries, cosmic messengers beyond light (i.e., gravitational waves, neutrinos, and cosmic rays) open entirely new windows to answer some of the most fundamental questions in (astro)physics: heavy element synthesis, equation of state of dense matter, particle acceleration, etc. This field was prioritized as a frontier scientific pursuit in the 2020 Decadal Survey on Astronomy and Astrophysics via its "New Windows on the Dynamic Universe" theme. MMA/TDA science presents technical challenges distinct from those experienced in other disciplines. Successful observations require coordination across myriad boundaries -- different cosmic messengers, ground vs. space, international borders, etc. -- all for sources that may not be well localized, and whose brightness may be changing rapidly with time. Add that all of this work is undertaken by real human beings, with distinct backgrounds, experiences, cultures, and expectations, that often conflict. To address these challenges and help MMA/TDA realize its full scientific potential in the coming decade (and beyond), the second in a series of community workshops sponsored by the U.S. National Science Foundation (NSF) and NASA titled "Windows on the Universe: Establishing the Infrastructure for a Collaborative Multi-Messenger Ecosystem" was held on October 16-18, 2023 in Tucson, AZ. Here we present the primary recommendations from this workshop focused on three key topics -- hardware, software, and people and policy. [abridged], Comment: Workshop white paper