Join us on November 9 to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers at the Low-Code/No-Code Summit. MLCommons is out today with its latest set of machine learning (ML) MLPerf benchmarks, once again showing how hardware and software for artificial intelligence (AI) are getting faster. MLCommons is a vendor-neutral organization that aims to provide standardized testing and benchmarks to help evaluate the state of ML software and hardware. Under the MLPerf testing name, MLCommons collects different ML benchmarks multiple times throughout the year. In September, the MLPerf Inference results were released, showing gains in how different technologies have improved inference performance.

Training machine learning (ML) algorithms is a computationally intensive process, which is frequently memory-bound due to repeatedly accessing large training datasets. As a result, processor-centric systems (e.g., CPU, GPU) suffer from costly data movement between memory units and processing units, which consumes large amounts of energy and execution cycles. Memory-centric computing systems, i.e., with processing-in-memory (PIM) capabilities, can alleviate this data movement bottleneck. Our goal is to understand the potential of modern general-purpose PIM architectures to accelerate ML training. To do so, we (1) implement several representative classic ML algorithms (namely, linear regression, logistic regression, decision tree, K-Means clustering) on a real-world general-purpose PIM architecture, (2) rigorously evaluate and characterize them in terms of accuracy, performance and scaling, and (3) compare to their counterpart implementations on CPU and GPU.

Davit Buniatyan is the Founding CEO at Activeloop, the company behind the fastest-growing dataset format specifically designed for AI. "If I want to tell you there is a spot on your shirt," Steve Jobs once said in an interview, "I'm not going to do it linguistically: 'There's a spot on your shirt 14 centimeters down from the collar and three centimeters to the left of your button.'" He would simply point at the spot. That was how he envisioned normal people using computers. While we realized this vision for day-to-day computer use, the same can't be said for working with data.

Davit Buniatyan is the Founding CEO at Activeloop, the company behind the fastest-growing dataset format specifically designed for AI. "If I want to tell you there is a spot on your shirt," Steve Jobs once said in an interview, "I'm not going to do it linguistically: 'There's a spot on your shirt 14 centimeters down from the collar and three centimeters to the left of your button.'" He would simply point at the spot. That was how he envisioned normal people using computers. While we realized this vision for day-to-day computer use, the same can't be said for working with data.

Serverless computing is a new type of cloud-based computation infrastructure initially developed for web microservices and IoT applications. As it frees model developers from concerns regarding capacity planning, configuration, management, maintenance, operating and scaling of containers, VMs and physical servers, serverless computing has gained popularity with machine learning (ML) researchers in recent years. Moreover, the benefits of serverless computing have also piqued interest in adopting it to data-intensive workloads such as ETL (extract, transform, load), query processing and ML, where it can provide significant cost reductions. Riding this trend, a research team from ETH Zürich and Microsoft recently conducted a systematic, comparative study of distributed ML training over serverless infrastructures (FaaS) and "serverful" infrastructures (IaaS), aiming to identify and understand the system tradeoffs involved in distributed ML training with serverless infrastructures. Serverless computing is offered by major cloud service providers such as AWS Lambda, Azure Functions and Google Cloud Functions.

The usability and practicality of any machine learning (ML) applications are largely influenced by two critical but hard-to-attain factors: low latency and low cost. Unfortunately, achieving low latency and low cost is very challenging when ML depends on real-world data that are highly distributed and rapidly growing (e.g., data collected by mobile phones and video cameras all over the world). Such real-world data pose many challenges in communication and computation. For example, when training data are distributed across data centers that span multiple continents, communication among data centers can easily overwhelm the limited wide-area network bandwidth, leading to prohibitively high latency and high cost. In this dissertation, we demonstrate that the latency and cost of ML on highly-distributed and rapidly-growing data can be improved by one to two orders of magnitude by designing ML systems that exploit the characteristics of ML algorithms, ML model structures, and ML training/serving data.

Machine-learning (ML) training is a beast of a problem. And it would appear that the industry is unleashing silicon beasts in order to battle that training beast. Some of those were revealed at the recent Hot Chips conference, including the biggest "chip" ever made. In honor of the benefits of being large, this article will be longer than average – and still leave things out. Most of what I've covered in the past has been about machine-learning inference.

The rise of industrial Internet of Things (IoT) and artificial intelligence (AI) are making edge computing significant for enterprises. Many industry verticals such as manufacturing, healthcare, automobile, transportation, and aviation are considering an investment in edge computing. Edge computing is fast becoming the conduit between the devices that generate data and the public cloud that processes the data. In the context of machine learning and artificial intelligence, the public cloud is used for training the models and the edge is utilized for inferencing. To accelerate ML training in the cloud, public cloud vendors such as AWS, Azure, and the Google Cloud Platform (GCP) offer GPU-backed virtual machines.

More than two billion images are shared daily in social networks alone. Research shows that it would take a person ten years to look at all the photos shared on Snapchat in the last hour! Media buyers and providers experience difficulty organizing relevant content in groups, parsing components of images/videos, and defining the return on investment from generated content in an efficient way. NVIDIA has many customers and ecosystem partners tackling that problem, using NVIDIA DGX as their preferred platform for deep learning (DL) powered image recognition. One of the notable names among the ecosystem is Imagga, a pioneer in offering a deep learning powered image recognition and image processing solution, built on NVIDIA DGX Station, the world's first personal AI supercomputer.

Abstract--Classifiers trained using conventional empirical risk minimization or maximum likelihood methods are known to suffer dramatic performance degradations when tested over examples adversarially selected based on knowledge of the classifier's decision rule. Due to the prominence of Artificial Neural Networks (ANNs) as classifiers, their sensitivity to adversarial examples, as well as robust training schemes, have been recently the subject of intense investigation. In this paper, for the first time, the sensitivity of spiking neural networks (SNNs), or third-generation neural networks, to adversarial examples is studied. The study considers rate and time encoding, as well as rate and first-to-spike decoding. Furthermore, a robust training mechanism is proposed that is demonstrated to enhance the performance of SNNs under white-box attacks.