Search

John Salmon,1 Daniel Bates,2 Neels Du Toit,2 Paul Verrills,2 James Yu,3 Murray G Taverner,4 Vahid Mohabbati,5 Matthew Green,6 Gary Heit7 ,†Robert Levy,8 Peter Staats,9 Shilpa Kottalgi,10 James Makous,11 Bruce Mitchell2 1Pain Management, Pain Care Perth, Perth Cottesloe, WA, Australia; 2Pain Management, Metro Pain Group, Melbourne, VIC, Australia; 3Pain Management, Sydney Spine and Pain, Sydney, NSW, Australia; 4Pain Management, Frankston Pain Management, Frankston, VIC, Australia; 5Pain Management, Sydney Pain Management Centre, Sydney, NSW, Australia; 6Pain Management, Pain Medicine of South Australia, Adelaide, SA, Australia; 7Department of Neurosurgery, Hue University of Medicine and Pharmacy, Hue, Vietnam; 8Neurosurgery, Institute for Neuromodulation, Boca Raton, FL, USA; 9Premier Pain Centers, Shrewsbury, NJ, USA; 10Clinical Department, Nalu Medical, Inc, Carlsbad, CA, USA; 11Makous Research, LLC, Carlsbad, CA, USA†Gary Heit passed away in February 2023Correspondence: John Salmon, Pain Management, Pain Care Perth, 2/89 Forrest Street, Perth Cottesloe, WA, 6011, Australia, Tel +61 428 246 846, Fax +61 8 92845759, Email Salmon8@bigpond.comPurpose: Spinal cord stimulation (SCS) is a highly effective treatment for chronic neuropathic pain. Despite recent advances in technology, treatment gaps remain. A small SCS system with a miniaturized implantable pulse generator (micro-IPG; < 1.5 cm3 in volume) and an externally worn power source may be preferred by patients who do not want a large, implanted battery. We report here the long-term outcomes from the first-in-human study evaluating the safety and performance of a new neurostimulation system.Patients and Methods: This was a prospective, multi-center, open-label, single-arm study to evaluate this SCS system, in the treatment of chronic, intractable leg and low-back pain. Consented subjects who passed screening continued on to the long-term phase of the study. One-year, patient-reported outcomes (PRO’s) suc

Numerical nonlinear algebra is a computational paradigm that uses numerical analysis to study polynomial equations. Its origins were methods to solve systems of polynomial equations based on the classical theorem of B\'ezout. This was decisively linked to modern developments in algebraic geometry by the polyhedral homotopy algorithm of Huber and Sturmfels, which exploits the combinatorial structure of the equations and led to efficient software for solving polynomial equations. Subsequent growth of numerical nonlinear algebra continues to be informed by algebraic geometry and its applications. These include new approaches to solving, algorithms for studying positive-dimensional varieties, certification, and a range of applications both within mathematics and from other disciplines. With new implementations, numerical nonlinear algebra is now a fundamental computational tool for algebraic geometry and its applications. We survey some of these innovations and some recent applications., Comment: 40 pages, many figures

We present a novel method for using Neural Networks (NNs) for finding solutions to a class of Partial Differential Equations (PDEs). Our method builds on recent advances in Neural Radiance Field research (NeRFs) and allows for a NN to converge to a PDE solution much faster than classic Physically Informed Neural Network (PINNs) approaches.

Muntjac is an open-source collection of components which can be used to build a multicore, Linux-capable system-on-chip. This includes a 64-bit RISC-V core, a cache subsystem, and TileLink interconnect allowing cache-coherent multicore configurations. Each component is easy to understand, verify, and extend, with most being configurable enough to be useful across a wide range of applications., Comment: To be published in the First Workshop on Open-Source Computer Architecture Research

Non-invasive epigenome editing is a promising strategy for engineering gene expression programs, yet potency, specificity, and persistence remain challenging. Here we show that effective epigenome editing is gated at single-base precision via 'keyhole' sites in endogenous regulatory DNA. Synthetic repressors targeting promoter keyholes can ablate gene expression in up to 99% of primary cells with single-gene specificity and can seamlessly repress multiple genes in combination. Transient exposure of primary T cells to keyhole repressors confers mitotically heritable silencing that persists to the limit of primary cultures in vitro and for at least 4 weeks in vivo, enabling manufacturing of cell products with enhanced therapeutic efficacy. DNA recognition and effector domains can be encoded as separate proteins that reassemble at keyhole sites and function with the same efficiency as single chain effectors, enabling gated control and rapid screening for novel functional domains that modulate endogenous gene expression patterns. Our results provide a powerful and exponentially flexible system for programming gene expression and therapeutic cell products.

The high energy costs of neural network training and inference led to the use of acceleration hardware such as GPUs and TPUs. While this enabled us to train large-scale neural networks in datacenters and deploy them on edge devices, the focus so far is on average-case performance. In this work, we introduce a novel threat vector against neural networks whose energy consumption or decision latency are critical. We show how adversaries can exploit carefully crafted $\boldsymbol{sponge}~\boldsymbol{examples}$, which are inputs designed to maximise energy consumption and latency. We mount two variants of this attack on established vision and language models, increasing energy consumption by a factor of 10 to 200. Our attacks can also be used to delay decisions where a network has critical real-time performance, such as in perception for autonomous vehicles. We demonstrate the portability of our malicious inputs across CPUs and a variety of hardware accelerator chips including GPUs, and an ASIC simulator. We conclude by proposing a defense strategy which mitigates our attack by shifting the analysis of energy consumption in hardware from an average-case to a worst-case perspective., Comment: Accepted at 6th IEEE European Symposium on Security and Privacy (EuroS&P)

Deep Graph Neural Networks (GNNs) show promising performance on a range of graph tasks, yet at present are costly to run and lack many of the optimisations applied to DNNs. We show, for the first time, how to systematically quantise GNNs with minimal or no loss in performance using Network Architecture Search (NAS). We define the possible quantisation search space of GNNs. The proposed novel NAS mechanism, named Low Precision Graph NAS (LPGNAS), constrains both architecture and quantisation choices to be differentiable. LPGNAS learns the optimal architecture coupled with the best quantisation strategy for different components in the GNN automatically using back-propagation in a single search round. On eight different datasets, solving the task of classifying unseen nodes in a graph, LPGNAS generates quantised models with significant reductions in both model and buffer sizes but with similar accuracy to manually designed networks and other NAS results. In particular, on the Pubmed dataset, LPGNAS shows a better size-accuracy Pareto frontier compared to seven other manual and searched baselines, offering a 2.3 times reduction in model size but a 0.4% increase in accuracy when compared to the best NAS competitor. Finally, from our collected quantisation statistics on a wide range of datasets, we suggest a W4A8 (4-bit weights, 8-bit activations) quantisation strategy might be the bottleneck for naive GNN quantisations.

Deep convolutional neural networks (CNNs) are powerful tools for a wide range of vision tasks, but the enormous amount of memory and compute resources required by CNNs pose a challenge in deploying them on constrained devices. Existing compression techniques, while excelling at reducing model sizes, struggle to be computationally friendly. In this paper, we attend to the statistical properties of sparse CNNs and present focused quantization, a novel quantization strategy based on power-of-two values, which exploits the weight distributions after fine-grained pruning. The proposed method dynamically discovers the most effective numerical representation for weights in layers with varying sparsities, significantly reducing model sizes. Multiplications in quantized CNNs are replaced with much cheaper bit-shift operations for efficient inference. Coupled with lossless encoding, we built a compression pipeline that provides CNNs with high compression ratios (CR), low computation cost and minimal loss in accuracy. In ResNet-50, we achieved a 18.08x CR with only 0.24% loss in top-5 accuracy, outperforming existing compression methods. We fully compressed a ResNet-18 and found that it is not only higher in CR and top-5 accuracy, but also more hardware efficient as it requires fewer logic gates to implement when compared to other state-of-the-art quantization methods assuming the same throughput., Comment: To appear in NeurIPS 2019, this is the same paper adapted for viewing on arXiv. TL;DR: Better size/accuracy trade-off of compressed sparse models with focused quantization. 11 pages, 5 figures, 4 tables

The Raspberry Pi is a single-board mini computer designed to get more people (particularly children) interested in computer programming. It aims to make programming tools and educational programs as accessible as possible, making it very easy to get started with. This book will guide you through six fun projects that show how programming can be used to be creative. Each project has clear step-by-step instructions and explanations helping children grasp the concepts easily. You will start by setting up the Raspberry Pi and get to grips with the Scratch programming language to create simple animations. Gain and put to use your understanding of Python to write simple yet useful programs. Create and play a physical game by connecting it to the Raspberry Pi en route to become aware of a number of other possible uses for a very similar circuit/program. Finally, with an understanding of Sonic Pi, you will create your own music.

The memory system of a modern embedded processor consumes a large fraction of total system energy. We explore a range of different configuration options and show that a reconfigurable design can make better use of the resources available to it than any fixed implementation, and provide large improvements in both performance and energy consumption. Reconfigurability becomes increasingly useful as resources become more constrained, so is particularly relevant in the embedded space. For an optimised architectural configuration, we show that a configurable cache system performs an average of 20% (maximum 70%) better than the best fixed implementation when two programs are competing for the same resources, and reduces cache miss rate by an average of 70% (maximum 90%). We then present a case study of AES encryption and decryption, and find that a custom memory configuration can almost double performance, with further benefits being achieved by specialising the task of each core when parallelising the program., Comment: Presented at HIP3ES, 2016

The Universal Accelerator Parser (UAP) is a library for reading and translating between lattice input formats. The UAP was primarily implemented to support the Accelerator Markup Language (AML) format and currently also supports the MAD lattice format. The UAP provides an extensible framework for reading and translating between different lattice formats. Included are routines for expression evaluation and beam line expansion. The use of a common library among accelerator codes will greatly improve the interoperability between different lattice file formats, and ease the development and maintenance to support these formats in programs. The UAP is written in C++ and compiles on most Unix, Linux, and Windows platforms a Java port is maintained for platform independence. Software developers can easily integrate the library into existing code by using the provided hooks.

We give a Descartes'-like bound on the number of positive solutions to a system of fewnomials that holds when its exponent vectors are not in convex position and a sign condition is satisfied. This was discovered while developing algorithms and software for computing the Gale transform of a fewnomial system, which is our main goal. This software is a component of a package we are developing for Khovanskii-Rolle continuation, which is a numerical algorithm to compute the real solutions to a system of fewnomials., Comment: 22 pages, 4 figures

This project resulted in software and algorithms for computing the real solution of polynomial systems contained within complex curves and surfaces. The primary product was Bertini Real, a publicly available software package for these computations. Secondary products included four articles, all in progress. These articles document the algorithms developed during the reporting period, including (1) a complete algorithm in the case of complex surfaces with no restricitons on the type of surface; (2) a technical article on the theory underlying and algorithm for finding critical points; (3) an article on Bertini Real; and (4) an article on the real numerical irreducible decomposition., Sponsored in part by DARPA.

To study the evolutionary dynamics of regulatory DNA, we mapped >1.3 million deoxyribonuclease I-hypersensitive sites (DHSs) in 45 mouse cell and tissue types, and systematically compared these with human DHS maps from orthologous compartments. We found that the mouse and human genomes have undergone extensive cis-regulatory rewiring that combines branch-specific evolutionary innovation and loss with widespread repurposing of conserved DHSs to alternative cell fates, and that this process is mediated by turnover of transcription factor (TF) recognition elements. Despite pervasive evolutionary remodeling of the location and content of individual cis-regulatory regions, within orthologous mouse and human cell types the global fraction of regulatory DNA bases encoding recognition sites for each TF has been strictly conserved. Our findings provide new insights into the evolutionary forces shaping mammalian regulatory DNA landscapes.

Numerical algebraic geometry is the field of computational mathematics concerning the numerical solution of polynomial systems of equations. Bertini, a popular software package for computational applications of this field, includes implementations of a variety of algorithms based on polynomial homotopy continuation. The Macaulay2 package Bertini.m2 provides an interface to Bertini, making it possible to access the core run modes of Bertini in Macaulay2. With these run modes, users can find approximate solutions to zero-dimensional systems and positive-dimensional systems, test numerically whether a point lies on a variety, sample numerically from a variety, and perform parameter homotopy runs.

Papers from the conference in honor of Andrew Sommese held at the University of Notre Dame, Notre Dame, IN, May 22--24, 2008. QC 20120216

A major problem in accelerator physics has been the sharing of lattice description files between programs. At present, there is no generally accepted lattice format that is comprehensive enough to meet current needs. To alleviate this problem, a lattice description format called Accelerator Markup Language (AML) has been created. AML is based upon the standard Extensible Markup Language (XML) format which gives AML the needed flexibility to be easily extended to changing requirements. In conjunction with AML, A set of parsing routines is being developed to speed the integration of AML into any program. These routines, collectively known as the Universal Accelerator Parser (UAP), will have the added feature of being able to translate MAD compatible lattice files and will also contain the needed hooks so that it will be a relatively simple matter to extend UAP to accommodate other lattice formats. Back parser routines to create AML and MAD lattice files will also be part of the UAP package.

Options being considered for the ILC Damping Rings include lattices with circumferences up to 17 km. The circumference, beam current and beam energy place the damping rings in a regime where resistive-wall instability is a concern, particularly as there are very demanding tolerances on the bunch-to-bunch jitter. Generally, it is possible to make good analytical estimates of the coupled-bunch growth rates in a storage ring, but particular features of the damping rings make it desirable to study the coupled-bunch instabilities using simulations. We present the results of simulations (including a bunch-by-bunch feedback system) of the transverse instabilities using a detailed lattice model.

We report results from studies of spin dynamics in the NLC Main Damping. Our studies have been based on spin tracking particles through the lattice under a range of conditions. We find that there are a number of spin resonances close to the nominal operating energy of 1.98 GeV; however, the effects of the resonances are weak, and the widths are narrow. We do not expect that any significant depolarization of the beam will occur during the store time.