Goto

Collaborating Authors

 chan


State Sen. Scott Wiener to succeed Pelosi in San Francisco House seat, as Chan concedes

Los Angeles Times

Things to Do in L.A. Tap to enable a layout that focuses on the article. State Sen. Scott Wiener to succeed Pelosi in San Francisco House seat, as Chan concedes State Sen. Scott Wiener speaks with supporters at his primary election night party at his San Francisco campaign headquarters on June 2. This is read by an automated voice. Please report any issues or inconsistencies here . See more from the L.A. Times in Google Search.


Josephine review: One of the most important films of the year is a hard, vital watch

Mashable

Look Up In My Bag Say More Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Mashable @Home Trending Now Mashable Selects Top creators, ranked Gift Ideas For Everyone On Your List Creator Playbook AI at School Safety Net All Series'Josephine' review: One of the most important films of the year is a hard, vital watch Shannon Connellan is Mashable's Senior Editor, General Assignments, based in London. She has been Mashable's UK Editor (and still manages the illustrious UK team) and Australia Editor, but emotionally, she lives searching for . A Tomatometer-approved critic, Shannon writes about entertainment, tech, social good, science, culture, and Australian horror, and loves to nerd out with movie stars, filmmakers, and TV creators . All products featured here are independently selected by our editors and writers. If you buy something through links on our site, Mashable may earn an affiliate commission.


Race to replace Pelosi enters the AI group chat

Los Angeles Times

Things to Do in L.A. Tap to enable a layout that focuses on the article. Sen. Scott Wiener, shown in 2023, caught heat for launching an AI chatbot parodying his rival in his race for congress. Scott Wiener's congressional campaign launched a chatbot parodying rival Connie Chan, sparking a rebuke from Pelosi and illustrating the minefield politicians navigate when using AI. This is read by an automated voice. Please report any issues or inconsistencies here .


China's dancing robots: how worried should we be?

The Guardian

Should we be impressed or worried by China's humanoid robot display? - video China's dancing robots: how worried should we be? Dancing humanoid robots took centre stage on Monday during the annual China Media Group's Spring Festival Gala, China's most-watched official television broadcast. They lunged and backflipped (landing on their knees), they spun around and jumped. The display was impressive, but prompted some to wonder: if robots can now dance and perform martial arts, what else can they do? Experts have mixed opinions, with some saying the robots had limitations and that the display should be viewed through a lens of state propaganda.




Octic Vision Transformers: Quicker ViTs Through Equivariance

arXiv.org Artificial Intelligence

Why are state-of-the-art Vision Transformers (ViTs) not designed to exploit natural geometric symmetries such as 90-degree rotations and reflections? In this paper, we argue that there is no fundamental reason, and what has been missing is an efficient implementation. To this end, we introduce Octic Vision Transformers (octic ViTs) which rely on octic group equivariance to capture these symmetries. In contrast to prior equivariant models that increase computational cost, our octic linear layers achieve 5.33x reductions in FLOPs and up to 8x reductions in memory compared to ordinary linear layers. In full octic ViT blocks the computational reductions approach the reductions in the linear layers with increased embedding dimension. We study two new families of ViTs, built from octic blocks, that are either fully octic equivariant or break equivariance in the last part of the network. Training octic ViTs supervised (DeiT-III) and unsupervised (DINOv2) on ImageNet-1K, we find that they match baseline accuracy while at the same time providing substantial efficiency gains.


Box, run, crash: China's humanoid robot games show advances and limitations

The Guardian

A quick left hook, a front kick to the chest, a few criss-cross jabs, and the crowd cheers. But it is not kickboxing prowess that concludes the match. It is an attempted roundhouse kick that squarely misses its target, sending the kickboxer from a top university team tumbling to the floor. While traditional kickboxing comes with the risk of blood, sweat and serious head injuries, the competitors in Friday's match at the inaugural World Humanoid Robot Games in Beijing faced a different set of challenges. The kickboxers, pint-sized humanoid robots entered by teams from leading Chinese technological universities, are part of a jamboree of humanoid events taking place at China's latest technology event.


A Gibbs Sampler for Efficient Bayesian Inference in Sign-Identified SVARs

arXiv.org Machine Learning

We develop a new algorithm for inference based on structural vector autoregressions (SVARs) identified with sign restrictions. The key insight of our algorithm is to break apart from the accept-reject tradition associated with sign-identified SVARs. We show that embedding an elliptical slice sampling within a Gibbs sampler approach can deliver dramatic gains in speed and turn previously infeasible applications into feasible ones. We provide a tractable example to illustrate the power of the elliptical slice sampling applied to sign-identified SVARs. We demonstrate the usefulness of our algorithm by applying it to a well-known small-SVAR model of the oil market featuring a tight identified set, as well as to a large SVAR model with more than 100 sign restrictions.


Statistical Mean Estimation with Coded Relayed Observations

arXiv.org Artificial Intelligence

We consider a problem of statistical mean estimation in which the samples are not observed directly, but are instead observed by a relay (``teacher'') that transmits information through a memoryless channel to the decoder (``student''), who then produces the final estimate. We consider the minimax estimation error in the large deviations regime, and establish achievable error exponents that are tight in broad regimes of the estimation accuracy and channel quality. In contrast, two natural baseline methods are shown to yield strictly suboptimal error exponents. We initially focus on Bernoulli sources and binary symmetric channels, and then generalize to sub-Gaussian and heavy-tailed settings along with arbitrary discrete memoryless channels.