Access to modern generative systems is often restricted to querying an API (the ``black-box" setting) and many properties of the system are unknown to the user at inference time. While recent work has shown that low-dimensional representations of models based on the relationship between their embedded responses to a set of queries are useful for inferring model-level properties, the quality of these representations is highly sensitive to the query set. We introduce the \emph{discriminative factorization} to distinguish between high- and low-quality query sets in the context of black-box model-level classification. Under this framework, the probability of chance-level classification decays exponentially in the query budget. On three auditing tasks, estimated factorization parameters predict the empirical performance decay rate. We conclude by showing that query sets selected using the estimated discriminative field reproduce the empirical ordering of oracle query sets.

Online learning from a stream of data is a defining feature of intelligence, yet modern machine learning systems often struggle in this setting, especially under distributional shift. To understand its basic properties, we study the relationship between online and offline learning in the context of kernel regression. We derive a closed-form expression for the function learned by online kernel regression, revealing that online kernel regression is equivalent to offline regression with shifted, inaccurate target outputs. Conversely, we show that by compensating for this effective shift in the teaching signal through target correction, online kernel-based learning can provably learn the same predictor as its offline counterpart. We derive both a closed-form expression for this target correction and an iterative form that can be applied sequentially. Applying this framework to image classification tasks on CIFAR-10 and CORe50, we show that online stochastic gradient descent with iteratively corrected targets outperforms learning with the true targets in continual learning settings. This work therefore provides a basic framework for analyzing and improving online learning in non-stationary environments.

Distributional treatment effects can be invisible to means: a treatment may preserve average outcomes while changing tails, modes, dispersion, or rare-event probabilities. Kernel tests can detect discrepancies between interventional outcome laws, but global tests do not reveal where the laws differ. We propose DR-ME, to our knowledge the first semiparametrically efficient finite-location test for interpretable distributional treatment effects. DR-ME evaluates an interventional kernel witness at learned outcome locations, returning causal-discrepancy coordinates rather than only a global rejection. From observational data, we derive orthogonal doubly robust kernel features whose centered oracle form is the canonical gradient of this finite witness. For fixed locations, we characterize the local testing limit: DR-ME is chi-square calibrated under the null, has noncentral chi-square local power, and uses the covariance whitening that optimizes local signal-to-noise for discrepancies visible through the selected coordinates. This efficient local-power geometry yields a principled location-learning criterion, with sample splitting preserving post-selection validity. Experiments show near-nominal type-I error, competitive power against global doubly robust kernel tests, and interpretable learned locations that localize distributional effects in a semi-synthetic medical-imaging study.

We study methods for simultaneous analysis of many noisy and biased estimates, each paired with an even noisier estimate of its own bias. The analyst's goal is to construct short calibrated intervals for each parameter. The standard debiasing approach, which subtracts the bias estimate from each biased estimate, inflates variance and yields long intervals. In this paper, we propose an empirical Bayes rebiasing strategy that starts from the fully debiased estimates and learns from data how much bias to reintroduce by estimating the unknown bias distribution. We provide convergence rates for the coverage of our intervals when the bias distribution is estimated using nonparametric maximum likelihood. Furthermore, we demonstrate substantial precision gains in prediction-powered inference, including pairwise LLM win-rate evaluations, as well as for inference of direct genetic effects in family-based GWAS.

Death of Alabama woman, 22, 'accidentally' shot in chest by boyfriend's dad is ruled a HOMICIDE Two small airlines join forces to create America's newest budget carrier after Spirit collapse leaves millions scrambling Horrifying final days of killer dad Chris Watts' pregnant wife before she was slaughtered alongside their daughters. Read all the chilling texts and receipts in full for first time: 'My eyes burn from crying' I'm a pastor who attended a secret UFO disclosure meeting. We saw images of'translucent beings' that chilled me to the bone... the files could fulfil a dark biblical prophecy Former NFL player Josh Mauro's tragic cause of death revealed after league was left'devastated' by ex-Cardinals and Giants man's sudden passing at 35 Cheerful Christian mom is pillar of Florida community and loves going on TV... but she has a childhood secret so evil that she stuttered with shock when confronted with it Taxpayers to foot Trump's $1.7 BILLION bill as President sues his own government: 'I'm paying myself' How I lost 3 STONE in 3 WEEKS. I've reversed pre-diabetes and no longer need a knee op: DONAL MACINTYRE's extraordinary investigation Popular megachurch in crisis as senior pastor suddenly quits... as bosses furiously DENY sex scandal Husband of doomed dive group leader says'something must have happened down there' as mystery surrounds why the five attempted to explore'cave so deep even divers with best equipment don't try' Greeks savage Kimberly Guilfoyle as Trump's ambassador opens McDonald's in country celebrated for world-class food Trump touts'fantastic' China trade win on Air Force One... but Wall Street is punishing the President I'm godfather to Candace Owens' daughter and Charlie Kirk was my friend... so I know the real reason she's attacking Erika - and I'll never publicly condemn her Wealthy dad'snarled the worst thing a parent could say' to younger daughter before he allegedly executed wife outside their gated community home during nightmare divorce Reese Witherspoon and Ryan Phillippe reunite for son's NYU graduation... as Kate Hudson cheers on her boy at same ceremony with Goldie Hawn and Kurt Russell'How do you live with that?' Disgraced Eric Swalwell's'blindsided' wife dresses for revenge... as friends reveal brutal toll sex assault scandal has had on young mom Judge declares another mistrial in disgraced Hollywood mogul Harvey Weinstein's rape case Can't lose weight no matter what you do? These are the 7 surprising reasons why, including'healthy' hacks actually making you put on pounds.

Samsung's Bespoke update is big step towards a useful AI for your fridge Samsung's Bespoke update is big step towards a useful AI for your fridge The idea of installing a software update on your fridge already feels kind of weird, let alone one centered around improving its AI capabilities. But that's exactly what's happening to Samsung's line of Bespoke refrigerators this week, and to my surprise this patch is making major strides at providing truly useful machine learning in a modern day icebox. As a quick recap, Samsung has offered AI-powered features like automatic food recognition and meal planning on its Bespoke refrigerators for a couple years already. However, as I found out after reviewing its flagship model late last year, the company's AI capabilities are still very much a work in progress. Previously, the fridge could recognize around 60 different kinds of fresh foods (like fruits and veggies) alongside another 50 or so packaged goods like yogurt or popcorn.

More information Adding us as a Preferred Source in Google by using this link indicates that you would like to see more of our content in Google News results. Simple could help ease your pain. Breakthroughs, discoveries, and DIY tips sent six days a week. Sciatica afflicts millions of people each year--though not as many people as they have it . A growing catchall term among the undiagnosed for all manner of back problems, sciatica is a specific lower-back nerve condition that requires specific action to address. "Early detection matters," John Gallucci Jr. MS, ATC, PT, DPT, the CEO of JAG Physical Therapy, tells .

Flytrex has partnered with Little Caesars to deliver full pizza orders by drone in Wylie, Texas, with food arriving in roughly four and a half minutes.

What are Russia's gains from the Iran war? 'We are not losers; we are winners' At least three people have been killed in Russian attacks on Ukraine in the past 24 hours despite a three-day ceasefire announced by US President Donald Trump that came into effect on May 9. Regional authorities on Sunday reported one death each in Ukraine's Zaporizhia, Dnipropetrovsk, and Kherson regions. Governor Oleksandr Prokudin confirmed the death on Telegram, saying the woman had been struck while walking down the street. Seven people, including a child, have also been injured across the region in drone or artillery attacks since early Saturday, he added. Ivan Fedorov, the governor of the southeastern Zaporizhia region, said one person had been killed and three others injured by artillery and drone attacks in the past 24 hours. In the northeastern Kharkiv region, Governor Oleh Syniehubov said eight people, including two children, were injured in drone attacks on the city of Kharkiv and nearby settlements.

I knew my writing students were using AI. It's what's lost when we surrender the struggle to translate thought into words I have been teaching fiction writing at MIT since 2017. Mark what works and what doesn't - underline great sentences, flag clunky syntax, gaps in logic and unrealistic dialogue. Ask yourself: does the story work? Answer in a signed letter to the author, attached to their story.