Congress apparently has strict limits on the use of ChatGPT and similar generative AI tools. Axios claims to have obtained a memo from House of Representatives administrative chief Catherine Szpindor setting narrow conditions for the use of ChatGPT and similar large language AI models in congressional offices. Staff are only allowed to use the paid ChatGPT Plus service due to its tighter privacy controls, and then only for "research and evaluation," Szpindor says. House offices are only allowed to use the chatbot with publicly accessible data even when using Plus, Szpindor adds. The privacy features have to be manually enabled to prevent interactions from feeding data into the AI model.

Everyone could use a longer battery life on their smartphone. CyberGuy shows you how to change your settings to make your battery last longer. Russian President Vladimir Putin is facing the biggest threat to his authority in decades. It comes after Yevgeny Prigozhin, who leads a private paramilitary group called Wagner, started a violent but brief uprising against Russia that may have long-lasting effects which could bring instability to Putin. If Putin is perceived as weakened by this assault on his leadership, then he is likely to exert aggressive strength in a show of power. CLICK TO GET KURT'S FREE CYBERGUY NEWSLETTER WITH SECURITY ALERTS, QUICK TIPS, TECH REVIEWS AND EASY HOW-TO'S TO MAKE YOU SMARTER What few know is that Vladimir Putin has been building a sophisticated facial recognition surveillance system since 2017 that is likely to be put into overdrive against any citizens willing to take on the views of the Russian leadership and military.

You can see your stress levels rising as the deadline to finish your memo approached, causing your beta brain wave activity to peak right before an alert popped up, telling you to take a brain break.

AI is defining the future, even as many US senators struggle to understand it in the present. "It would have been better if it had been held in a room where the acoustics were better," Senator Chuck Grassley, an Iowa Republican, says of a much-anticipated--if overdue--All-Senators AI briefing orchestrated by Senate Majority Leader Chuck Schumer earlier this month. The shoddy acoustics of the first of three closed-door meetings--kept private to insulate senators from electoral pressure to perform before cameras--were far from Grassley's biggest complaint. "I would say that the next [one] will be more valuable, because this was a very general overview," he says. As AI expands its foothold across industries, households, and legislative bodies--including amongst some at the Capitol itself--Congress is under pressure to act quickly, even though many lawmakers still don't know what they're being asked to regulate.

Despite our intrinsic distrust of AI in space taught to us by movies like 2001: A Space Odyssey ("I'm afraid I can't do that, Dave"), it offers large advantages to both manned and unmanned missions. To that end, NASA is developing a system that will allow astronauts to perform maneuvers, conduct experiments and more using a natural-language ChatGPT-like interface, The Guardian reported. "The idea is to get to a point where we have conversational interactions with space vehicles and they [are] also talking back to us on alerts, interesting findings they see in the solar system and beyond," said Dr. Larissa Suzuki, speaking at an IEEE meeting on next-gen space communication. NASA aims to deploy the system on its Lunar Gateway, a space station that will orbit the Moon and provide support for NASA's Artemis mission. It would use a natural language interface that allows astronauts to seek advice on experiments or conduct maneuvers without diving into complex manuals. On a dedicated page soliciting small business support for Lunar Gateway, NASA wrote that it would require AI and machine learning technologies to manage various systems when it's unoccupied as well.

Navy SEAL Jake Zweig responds to the intense search for the missing Titanic submarine on'Fox & Friends.' Efforts to recover the remains of the Titan submersible that suffered a catastrophic implosion near the Titanic wreckage are currently underway, and as of Sunday, had descended to the seafloor for a fourth dive. Last Thursday, the U.S. Coast Guard confirmed that a debris field located about 1,600 feet from the wreckage of the Titanic was in fact that of the missing Titan submersible. The underwater vessel was carrying five men on board when it lost contact with its surface ship about an hour and 45 minutes after descending to the Titanic. South Wellfleet, Massachusetts-based Pelagic Research Services (PRS) was contacted by OceanGate, the company behind Titan, for use of its remotely operated vehicles, or "ROVs," to assist with the search. Pelagic Research Services continues to assist the Transportation Safety Board of Canada, U.S. Coast Guard, and U.S. National Transportation Safety Board with Titan recovery efforts near the Titanic wreckage.

The reliable operation of automatic systems is heavily dependent on the ability to detect faults in the underlying dynamical system. While traditional model-based methods have been widely used for fault detection, data-driven approaches have garnered increasing attention due to their ease of deployment and minimal need for expert knowledge. In this paper, we present a novel principal component analysis (PCA) method that uses occupation kernels. Occupation kernels result in feature maps that are tailored to the measured data, have inherent noise-robustness due to the use of integration, and can utilize irregularly sampled system trajectories of variable lengths for PCA. The occupation kernel PCA method is used to develop a reconstruction error approach to fault detection and its efficacy is validated using numerical simulations.

Adversarial training is one of the best-performing methods in improving the robustness of deep language models. However, robust models come at the cost of high time consumption, as they require multi-step gradient ascents or word substitutions to obtain adversarial samples. In addition, these generated samples are deficient in grammatical quality and semantic consistency, which impairs the effectiveness of adversarial training. To address these problems, we introduce a novel, effective procedure for instead adversarial training with only clean data. Our procedure, distribution shift risk minimization (DSRM), estimates the adversarial loss by perturbing the input data's probability distribution rather than their embeddings. This formulation results in a robust model that minimizes the expected global loss under adversarial attacks. Our approach requires zero adversarial samples for training and reduces time consumption by up to 70\% compared to current best-performing adversarial training methods. Experiments demonstrate that DSRM considerably improves BERT's resistance to textual adversarial attacks and achieves state-of-the-art robust accuracy on various benchmarks.

In-context learning (ICL) improves language models' performance on a variety of NLP tasks by simply demonstrating a handful of examples at inference time. It is not well understood why ICL ability emerges, as the model has never been specifically trained on such demonstrations. Unlike prior work that explores implicit mechanisms behind ICL, we study ICL via investigating the pretraining data. Specifically, we first adapt an iterative, gradient-based approach to find a small subset of pretraining data that supports ICL. We observe that a continued pretraining on this small subset significantly improves the model's ICL ability, by up to 18%. We then compare the supportive subset constrastively with random subsets of pretraining data and discover: (1) The supportive pretraining data to ICL do not have a higher domain relevance to downstream tasks. (2) The supportive pretraining data have a higher mass of rarely occurring, long-tail tokens. (3) The supportive pretraining data are challenging examples where the information gain from long-range context is below average, indicating learning to incorporate difficult long-range context encourages ICL. Our work takes a first step towards understanding ICL via analyzing instance-level pretraining data. Our insights have a potential to enhance the ICL ability of language models by actively guiding the construction of pretraining data in the future.

Differential Privacy (DP) ensures that training a machine learning model does not leak private data. However, the cost of DP is lower model accuracy or higher sample complexity. In practice, we may have access to auxiliary public data that is free of privacy concerns. This has motivated the recent study of what role public data might play in improving the accuracy of DP models. In this work, we assume access to a given amount of public data and settle the following fundamental open questions: 1. What is the optimal (worst-case) error of a DP model trained over a private data set while having access to side public data? What algorithms are optimal? 2. How can we harness public data to improve DP model training in practice? We consider these questions in both the local and central models of DP. To answer the first question, we prove tight (up to constant factors) lower and upper bounds that characterize the optimal error rates of three fundamental problems: mean estimation, empirical risk minimization, and stochastic convex optimization. We prove that public data reduces the sample complexity of DP model training. Perhaps surprisingly, we show that the optimal error rates can be attained (up to constants) by either discarding private data and training a public model, or treating public data like it's private data and using an optimal DP algorithm. To address the second question, we develop novel algorithms which are "even more optimal" (i.e. better constants) than the asymptotically optimal approaches described above. For local DP mean estimation with public data, our algorithm is optimal including constants. Empirically, our algorithms show benefits over existing approaches for DP model training with side access to public data.