Fox News senior national correspondent William La Jeunesse reports on proposed changes to California's electric bills on'Special Report.' Power saver programs are utility-sponsored demand response initiatives that help reduce electricity usage during periods of peak demand. These programs typically target central air conditioners (AC) and heat pumps, since cooling equipment drives spikes in summer energy demand. In exchange for incentives such as bill credits or rebates, participating homeowners allow the utility to temporarily adjust or cycle their HVAC systems on hot days. I recently received an email from Leah, an HVAC professional based in Rio Rancho, New Mexico.

Efficient multi-hop reasoning requires Large Language Models (LLMs) based agents to acquire high-value external knowledge iteratively. Previous work has explored reinforcement learning (RL) to train LLMs to perform search-based document retrieval, achieving notable improvements in QA performance, but underperform on complex, multi-hop QA resulting from the sparse rewards from global signal only. To address this gap in existing research, we introduce StepSearch, a framework for search LLMs that trained with step-wise proximal policy optimization method. It consists of richer and more detailed intermediate search rewards and token-level process supervision based on information gain and redundancy penalties to better guide each search step. We constructed a fine-grained question-answering dataset containing sub-question-level search trajectories based on open source datasets through a set of data pipeline method. On standard multi-hop QA benchmarks, it significantly outperforms global-reward baselines, achieving 11.2% and 4.2% absolute improvements for 3B and 7B models over various search with RL baselines using only 19k training data, demonstrating the effectiveness of fine-grained, stepwise supervision in optimizing deep search LLMs. Our code will be released on https://github.com/Zillwang/StepSearch.

Visually-situated language is ubiquitous -- sources range from textbooks with diagrams to web pages with images and tables, to mobile apps with buttons and forms. Perhaps due to this diversity, previous work has typically relied on domain-specific recipes with limited sharing of the underlying data, model architectures, and objectives. We present Pix2Struct, a pretrained image-to-text model for purely visual language understanding, which can be finetuned on tasks containing visually-situated language. Pix2Struct is pretrained by learning to parse masked screenshots of web pages into simplified HTML. The web, with its richness of visual elements cleanly reflected in the HTML structure, provides a large source of pretraining data well suited to the diversity of downstream tasks. Intuitively, this objective subsumes common pretraining signals such as OCR, language modeling, image captioning. In addition to the novel pretraining strategy, we introduce a variable-resolution input representation and a more flexible integration of language and vision inputs, where language prompts such as questions are rendered directly on top of the input image. For the first time, we show that a single pretrained model can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images.

In June, the school board in Rio Rancho, New Mexico, was facing a series of votes on the budget for an elaborate and expensive reopening plan. Among the big-ticket items was a tablet designed to screen students and staff for fevers. The devices were sold by a company named OneScreen, which supplies schools with technology including "smart" whiteboards and attendance apps. But this spring, it had pivoted. Its new product, called GoSafe, could scan foreheads for elevated temperatures and detect when students aren't wearing masks.

Symposia will be limited to between forty and sixty participants. Each participant will be expected to attend a single symposium. In addition to invited participants, a limited number of other interested parties will be allowed to register in each symposium on a first-come, first-served basis. Working notes will be prepared and distributed to participants in each symposium, but will not otherwise be available unless published as an AAAI Technical Report or edited collection. The final deadline for registration is October 12, 2007. For registration information, please contact AAAI at fss07@aaai.org or visit AAAI's web site (www.aaai.org/Symposia/Fall/fss07.