One chilly day in November 2025, community worker Mike Prado drove through Abilene, Tex., handing out blankets, socks, and jackets to unhoused individuals across the city. People sat on curbs, alleyway after alleyway, their meager belongings soaked by the previous night's hard rain. Prado has worked in this community for a decade, and was once homeless in Abilene himself. Prado has witnessed difficult years--but the current situation was the worst he'd ever seen, he told TIME. One man with a walker approached Prado outside of the Hope Haven offices--an Abilene nonprofit where Prado works, which operates a shelter and helps people with vouchers find housing--and accepted a jacket from him.

Unfortunately, exact computations for both are NP-hard.

Diffusion models can be improved with additional guidance towards more effective representations of input. Indeed, prior empirical work has already shown that aligning internal representations of the diffusion model with those of pre-trained models improves generation quality. In this paper, we present a systematic framework for incorporating representation guidance into diffusion models. We provide alternative decompositions of denoising models along with their associated training criteria, where the decompositions determine when and how the auxiliary representations are incorporated. Guided by our theoretical insights, we introduce two new strategies for enhancing representation alignment in diffusion models. First, we pair examples with target representations either derived from themselves or arisen from different synthetic modalities, and subsequently learn a joint model over the multimodal pairs. Second, we design an optimal training curriculum that balances representation learning and data generation. Our experiments across image, protein sequence, and molecule generation tasks demonstrate superior performance as well as accelerated training. In particular, on the class-conditional ImageNet $256\times 256$ benchmark, our guidance results in $23.3$ times faster training than the original SiT-XL as well as four times speedup over the state-of-the-art method REPA. The code is available at https://github.com/ChenyuWang-Monica/REED.

The number of planning approvals for new homes in England is unacceptable, the new housing secretary has said, after official data showed permission for building homes fell to a record low during Labour's first year in office. Fewer than 29,000 projects were granted permission by councils in the year ending June 2025 - striking a blow to the government's promise to deliver 1.5 million homes by the next election. Steve Reed, who has taken over from Angela Rayner as housing secretary, said fixing the planning system won't happen overnight. Conservative shadow housing secretary Sir James Cleverly said that Labour had promised to'build, build, build' but their flagship planning reforms clearly aren't working. You can see the figures for your local area in BBC Verify's housing tracker.

DeeperDive, a new tool that converses with readers, is an effort to beat the AI industry at its own game. The publishing company behind USA Today and 220 other publications is today rolling out a chatbot -like tool called DeeperDive that can converse with readers, summarize insights from its journalism, and suggest new content from across its sites. "Visitors now have a trusted AI answer engine on our platform for anything they want to engage with, anything they want to ask," Mike Reed, CEO of Gannett and the USA Today Network, said at the WIRED AI Power Summit in New York, an event that brought together voices from the tech industry, politics, and the world of media. "and it is performing really great." Most publishers have a fraught relationship with AI, as the chatbots that trained on their content are now summarizing it and eating the traffic that search engines used to send them.

Example 1 Figure 1 shows a graph mapping. Edge mappings can be trivially inferred. Hence, the claim is proved. These four cases cover all possible situations and hence, the triangle inequality is established. From the triangle inequality, we can infer the lower bounds listed in lines 2 and 4 of Alg. 2. Hence, if Alg. 3 presents the pseudocode.

Former deputy chief of staff for policy Bruce Reed arrived on Capitol Hill for his closed-door deposition with the House Oversight Committee on Tuesday. Former White House physician Kevin O'Connor previously dismissed cognitive tests as "meaningless," ex-Biden administration aide Bruce Reed told House investigators on Tuesday, according to a source familiar with the proceedings. Reed, who served as White House deputy chief of staff for policy, is the ninth member of former President Joe Biden's inner circle to sit down with House Oversight Committee lawyers. A source familiar with his interview told Fox News Digital that Reed attributed Biden's disastrous 2024 debate performance against then-candidate Donald Trump to the former president's stutter, a condition that's been well-documented and Biden himself has publicly acknowledged. But his meandering and seemingly tired demeanor on stage with Trump alarmed both Democrats and media pundits, who saw it as a glaring sign of Biden's advanced age.

While existing benchmarks probe the reasoning abilities of large language models (LLMs) across diverse domains, they predominantly assess passive reasoning, providing models with all the information needed to reach a solution. By contrast, active reasoning-where an LLM must interact with external systems to acquire missing evidence or data-has received little systematic attention. To address this shortfall, we present AR-Bench, a novel benchmark designed explicitly to evaluate an LLM's active reasoning skills. AR-Bench comprises three task families-detective cases, situation puzzles, and guessing numbers-that together simulate real-world, agentic scenarios and measure performance across commonsense, logical, and symbolic reasoning challenges. Empirical evaluation on AR-Bench demonstrates that contemporary LLMs exhibit pronounced difficulties with active reasoning: they frequently fail to acquire or leverage the information needed to solve tasks. This gap highlights a stark divergence between their passive and active reasoning abilities. Moreover, ablation studies indicate that even advanced strategies, such as tree-based searching or post-training approaches, yield only modest gains and fall short of the levels required for real-world deployment. Collectively, these findings highlight the critical need to advance methodology for active reasoning, e.g., incorporating interactive learning, real-time feedback loops, and environment-aware objectives for training. The benchmark is publicly available at: https://github.com/tmlr-group/AR-Bench.