Point cloud architecture design has become a crucial problem for 3D deep learning. Several efforts exist to manually design architectures with high accuracy in point cloud tasks such as classification, segmentation, and detection. Recent progress in automatic Neural Architecture Search (NAS) minimizes the human effort in network design and optimizes high performing architectures. However, these efforts fail to consider important factors such as latency during inference. Latency is of high importance in time critical applications like self-driving cars, robot navigation, and mobile applications, that are generally bound by the available hardware. In this paper, we introduce a new NAS framework, dubbed LC-NAS, where we search for point cloud architectures that are constrained to a target latency. We implement a novel latency constraint formulation to trade-off between accuracy and latency in our architecture search. Contrary to previous works, our latency loss guarantees that the final network achieves latency under a specified target value. This is crucial when the end task is to be deployed in a limited hardware setting. Extensive experiments show that LC-NAS is able to find state-of-the-art architectures for point cloud classification in ModelNet40 with minimal computational cost. We also show how our searched architectures achieve any desired latency with a reasonably low drop in accuracy. Finally, we show how our searched architectures easily transfer to a different task, part segmentation on PartNet, where we achieve state-of-the-art results while lowering latency by a factor of 10.
Brazilians have seen recent improvements in fixed broadband latency as demand for online gaming rises during the Covid-19 outbreak, a new study has found. Latency - the reaction time of a connection - varies between countries across Latin America, particularly when it comes to fixed broadband. Latency is a key metric in gaming and determines much of the user's experience in terms of the absence of lags during gameplay. According to the data from Ookla's Speedtest Intelligence, gamers in Brazil had the lowest mean latency on fixed broadband, relevant for games played on PC and console games, at 19 ms during Q2 2020, down from 23 ms in the same period in 2019. By comparison, Colombia had the highest fixed broadband latency at 43 ms.
A new Accenture survey of nearly 2,000 technology and business executives in 10 countries revealed deep uncertainty about next-generation mobile network technology, known as 5G. Few of those surveyed, for example, believe industry predictions about the dramatically improved speeds of 5G networks. And more than half don't expect the technology will enable them to do much that they can't already do. Nearly three-quarters said they need help imagining 5G use cases. These findings suggest that many business leaders neither understand the technology nor its disruptive potential.
Google on Wednesday took the wraps off latest version of its Andromeda, the a software-defined network (SDN) stack that underpins the Google Cloud Platform (GCP). With this release, packets on the network are able to securely bypass the hypervisor and go directly to the Andromeda software packet processor. This should reduce latency between Compute Engine VMs by 40 percent over the previous version of Andromeda. Since Google launched Andromeda in 2014, it's been able to reduce Compute Engine latency by nearly a factor of eight. "Without having to do anything, [customers'] code in some cases will run noticeably faster and better," Google engineering fellow Amin Vadhat told ZDNet.
After a few weeks of public testing on the PTR client, Overwatch is the latest game to get full support for NVIDIA's low-latency Reflex technology on PC. Reflex will go live on Overwatch's regular servers today alongside the game's latest seasonal event. Reflex is all about reducing system latency, the time between clicking your mouse and seeing the corresponding action on your screen. It works by slashing the number of frames that your GPU lines up in certain circumstances, which eases the strain on your CPU. NVIDIA says the tech can reduce input lag by up to 50 percent, with less-powerful GPUs getting a bigger boost than the likes of GeForce 30-series cards.