Chain-of-Thought (CoT) prompting is a key technique for enabling complex reasoning in large language models. However, generating full, fixed-length rationales is computationally wasteful, inflating both token usage and latency. We introduce LEASH: Logit-Entropy Adaptive Stopping Heuristic, a training-free decoding algorithm that adaptively halts rationale generation. LEASH monitors two intrinsic signals: the slope of token-level entropy and the improvement in the top-logit margin. It terminates the generation once both signals plateau, indicating the model has reached a stable reasoning state. Across four instruction-tuned models on the GSM8K and AQuA-RAT benchmarks, LEASH reduces average token generation by 30--35% and latency by 27%, while incurring a 10 p.p. accuracy drop relative to CoT. LEASH is model-agnostic and requires no additional training or supervision, offering a simple and efficient alternative to CoT decoding.

An NYPD drone observed four minors, between the ages of 12 and 16 years old, riding on top of a train in the Bronx on Thursday as it passed multiple stations at a high speed. FIRST ON FOX: A new move by Congress would unleash civilian drone use across America's skies by establishing rules to allow them to be flown beyond a user's line of sight and using AI for approval to do so. Her LIFT Act, introduced in the House on Thursday, would require Transportation Secretary Sean Duffy to establish set performance and safety standards for BVLOS operations and review current aviation standards, which were designed with manned aircraft in mind. It would also require the Transportation secretary to deploy artificial intelligence to assist with processing waiver applications to allow civilian drones to fly BVLOS. Industry operators have long pushed for new BVLOS policy to replace the current system in which individuals must apply for waivers with the Federal Aviation Adminsitration (FAA) through a costly, cumbersome process to fly beyond the line of sight.

This paper investigates the mathematical reasoning capabilities of large language models (LLMs) using 50 newly constructed high-school-level word problems. Unlike prior studies that focus solely on answer correctness, we rigorously analyze both final answers and solution steps to identify reasoning failures. Evaluating eight state-of-the-art models - including Mixtral, Llama, Gemini, GPT-4o, and OpenAI's o1 variants - we find that while newer models (e.g., o3-mini, deepseek-r1) achieve higher accuracy, all models exhibit errors in spatial reasoning, strategic planning, and arithmetic, sometimes producing correct answers through flawed logic. Common failure modes include unwarranted assumptions, over-reliance on numerical patterns, and difficulty translating physical intuition into mathematical steps. Manual analysis reveals that models struggle with problems requiring multi-step deduction or real-world knowledge, despite possessing broad mathematical knowledge. Our results underscore the importance of evaluating reasoning processes, not just answers, and caution against overestimating LLMs' problem-solving proficiency. The study highlights persistent gaps in LLMs' generalization abilities, emphasizing the need for targeted improvements in structured reasoning and constraint handling.

Robotic guide dogs hold significant potential to enhance the autonomy and mobility of blind or visually impaired (BVI) individuals by offering universal assistance over unstructured terrains at affordable costs. However, the design of robotic guide dogs remains underexplored, particularly in systematic aspects such as gait controllers, navigation behaviors, interaction methods, and verbal explanations. Our study addresses this gap by conducting user studies with 18 BVI participants, comprising 15 cane users and three guide dog users. Participants interacted with a quadrupedal robot and provided both quantitative and qualitative feedback. Our study revealed several design implications, such as a preference for a learning-based controller and a rigid handle, gradual turns with asymmetric speeds, semantic communication methods, and explainability. The study also highlighted the importance of customization to support users with diverse backgrounds and preferences, along with practical concerns such as battery life, maintenance, and weather issues. These findings offer valuable insights and design implications for future research and development of robotic guide dogs.

Navigation in cluttered underwater environments is challenging, especially when there are constraints on communication and self-localisation. Part of the fully distributed underwater navigation problem has been resolved by introducing multi-agent robot teams, however when the environment becomes cluttered, the problem remains unresolved. In this paper, we first studied the connection between everyday activity of dog walking and the cooperative underwater obstacle avoidance problem. Inspired by this analogy, we propose a novel dog walking paradigm and implement it in a multi-agent underwater system. Simulations were conducted across various scenarios, with performance benchmarked against traditional methods utilising Image-Based Visual Servoing in a multi-agent setup. Results indicate that our dog walking-inspired paradigm significantly enhances cooperative behavior among agents and outperforms the existing approach in navigating through obstacles.

But people who tried it out this past week found that the tool, built on the popular ChatGPT system, could quickly veer into strange territory. It showed signs of defensiveness over its name with a Washington Post reporter and told a New York Times columnist it wanted to break up his marriage. It also claimed an Associated Press reporter was "being compared to Hitler because you are one of the most evil and worst people in history."

Robots that are designed to act like dogs are becoming more common on Chinese social media, but it remains to be seen if they will catch on.

"Why AI is harder than we think" - that's the title of a recent paper by Melanie Mitchell at the Santa Fe Institute. The paper is in two parts. The first explores the history of AI, beginning with McCarthy in 1957 to the present. She explains that the AI process has been instead punctuated, and she explains in a way that seems plausible. AI always seems to burst out with dramatic advances, then doesn't live up to the hype, and hibernates for a few years, the so-called AI Winter.

If you make a purchase by clicking one of our links, we may earn a small share of the revenue. However, our picks and opinions are independent from USA TODAY's newsroom and any business incentives. Don't scoff: smart products make life so much easier. Consider that your Amazon Alexa can help you cook, your Google Home can improve your morning routine, and your iPhone can take a better photo than a fancy camera. Why shouldn't we apply smart technology to our furry, beloved family members?

Here's the most critical thing for CEOs to understand about machine learning: When it's done right, the outcomes get better with more time, more data, more use. That's so important because companies that aren't using machine learning to, say, make better product recommendations or provide better customer service risk falling behind rivals whose systems keep getting smarter. "Every CEO, in every industry, needs to worry about new companies coming in and doing it better by using software," says Amit Zavery, senior vice president for Oracle Cloud Platform. "The next thing most companies are going to do is start using machine learning to improve the software--and improve their relationships with customers. So it's going to be too late if you don't learn now."