Linear Temporal Logic (LTL) has become a prevalent specification language for robotic tasks. To mitigate the significant manual effort and expertise required to define LTL-encoded tasks, several methods have been proposed for translating Natural Language (NL) instructions into LTL formulas, which, however, lack correctness guarantees. To address this, we introduce a new NL-to-LTL translation method, called ConformalNL2LTL, that can achieve user-defined translation success rates over unseen NL commands. Our method constructs LTL formulas iteratively by addressing a sequence of open-vocabulary Question-Answering (QA) problems with LLMs. To enable uncertainty-aware translation, we leverage conformal prediction (CP), a distribution-free uncertainty quantification tool for black-box models. CP enables our method to assess the uncertainty in LLM-generated answers, allowing it to proceed with translation when sufficiently confident and request help otherwise. We provide both theoretical and empirical results demonstrating that ConformalNL2LTL achieves user-specified translation accuracy while minimizing help rates.

--Recently, Large Language Models (LLMs) have demonstrated remarkable advancements in Natural Language Processing (NLP). However, generating high-quality text that balances coherence, diversity, and relevance remain s challenging. Traditional decoding methods, such as bean search and top-k sampling, often struggle with either repe titive or incoherent outputs, particularly in tasks that require long-form text generation. To address these limitations, the paper proposes a novel enhancement of the well-known Contrastive S earch algorithm, Context-Enhanced Contrastive Search (CEC S) with contextual calibration. The proposed scheme introduces several novelties including dynamic contextual importance w eighting, multi-level Contrastive Search, and adaptive temper ature control, to optimize the balance between fluency, creativity, and precision. The performance of CECS is evaluated usi ng several standard metrics such as BLEU, ROUGE, and semantic similarity. Experimental results demonstrate signif icant improvements in both coherence and relevance of the generated texts by CECS outperforming the existing Contrastive Search techniques. The proposed algorithm has several pote ntial applications in the real world including legal document drafting, customer service chatbots, and content marketing. In recent years, Large Language Models (LLMs) have transformed the field of Natural Language Processing (NLP), delivering cutting-edge performance across numerous tasks, including text generation, summarization, machine translation, and question answering. Models such as OpenAI's GPT-3 [1], Google's BERT [2], and more recently PaLM [3], have greatly enhanced the capabilities of machines in understanding and generating human language. By leveraging deep neural network architectures and training on extensive datasets, LLMs have made significant strides in pro ducing fluent and coherent text that closely resembles hum an communication. Generating text from an LLM involves more than simp ly predicting the next word in a sequence according to its probability distribution. This step, known as decod ing, plays a critical role in shaping the final output. Various decoding strategies have been proposed in the literature ranging from deterministic methods such as beam search, to stoch astic methods like top-k and nucleus sampling. While the deterministic methods choose the highest probability token at each step, their stochastic counterparts introduce randomness to improve diversity in the generated output.

The rapid advancement of Large Language Models (LLMs) has driven the need for more efficient serving strategies. In this context, efficiency refers to the proportion of requests that meet their Service Level Objectives (SLOs), particularly for Time To First Token (TTFT) and Time Between Tokens (TBT). However, existing systems often prioritize one metric at the cost of the other. We present Ascendra, an LLM serving system designed to meet both TTFT and TBT SLOs simultaneously. The core insight behind Ascendra is that a request's urgency evolves as it approaches its deadline. To leverage this, Ascendra partitions GPU resources into two types of instances: low-priority and high-priority. Low-priority instances maximize throughput by processing requests out of arrival order, but at the risk of request starvation. To address this, Ascendra employs a performance model to predict requests at risk of missing their SLOs and proactively offloads them to high-priority instances. High-priority instances are optimized for low-latency execution and handle urgent requests nearing their deadlines. This partitioned architecture enables Ascendra to effectively balance high throughput and low latency. Extensive evaluation shows that Ascendra improves system throughput by up to 1.7x compared to vLLM and Sarathi-Serve while meeting both TTFT and TBT SLOs.

Retrieval-augmented generation (RAG) systems face significant challenges in multi-hop question answering (MHQA), where complex queries require synthesizing information across multiple document chunks. Existing approaches typically rely on iterative LLM-based query rewriting and routing, resulting in high computational costs due to repeated LLM invocations and multi-stage processes. To address these limitations, we propose TreeHop, an embedding-level framework without the need for LLMs in query refinement. TreeHop dynamically updates query embeddings by fusing semantic information from prior queries and retrieved documents, enabling iterative retrieval through embedding-space operations alone. This method replaces the traditional "Retrieve-Rewrite-Vectorize-Retrieve" cycle with a streamlined "Retrieve-Embed-Retrieve" loop, significantly reducing computational overhead. Moreover, a rule-based stop criterion is introduced to further prune redundant retrievals, balancing efficiency and recall rate. Experimental results show that TreeHop rivals advanced RAG methods across three open-domain MHQA datasets, achieving comparable performance with only 5\%-0.4\% of the model parameter size and reducing the query latency by approximately 99\% compared to concurrent approaches. This makes TreeHop a faster and more cost-effective solution for deployment in a range of knowledge-intensive applications. For reproducibility purposes, codes and data are available here: https://github.com/allen-li1231/TreeHop-RAG.

Large Language Models (LLMs) have transformed software engineering, but their application to physical engineering domains remains underexplored. This paper evaluates LLMs' capabilities in high-powered rocketry design through RocketBench, a benchmark connecting LLMs to high-fidelity rocket simulations. We test models on two increasingly complex design tasks: target altitude optimization and precision landing challenges. Our findings reveal that while state-of-the-art LLMs demonstrate strong baseline engineering knowledge, they struggle to iterate on their designs when given simulation results and ultimately plateau below human performance levels. However, when enhanced with reinforcement learning (RL), we show that a 7B parameter model outperforms both SoTA foundation models and human experts. This research demonstrates that RL-trained LLMs can serve as effective tools for complex engineering optimization, potentially transforming engineering domains beyond software development.

Large Language Models (LLMs) are powerful but often require extensive fine-tuning and large datasets for specialized domains like law. General-purpose pre-training may not capture legal nuances, and acquiring sufficient legal data is challenging. We introduce SynLexLM, a novel approach to efficiently pre-train a legal LLM. Our method employs curriculum learning, progressing from simple to complex legal texts and queries, combined with synthetic data augmentation using models like Gemini Pro to address data scarcity. We aim to achieve improved performance on legal benchmarks (BigLaw-Bench, EUR-Lex-Sum) compared to traditional models and fine-tuned versions. Preliminary work involves generating synthetic QA pairs reflecting legal reasoning. This work aims to enhance legal document analysis and research tools, potentially democratizing access to advanced legal AI.

Current multimodal benchmarks often conflate reasoning with domain-specific knowledge, making it difficult to isolate and evaluate general reasoning abilities in non-expert settings. To address this, we introduce VisualPuzzles, a benchmark that targets visual reasoning while deliberately minimizing reliance on specialized knowledge. VisualPuzzles consists of diverse questions spanning five categories: algorithmic, analogical, deductive, inductive, and spatial reasoning. One major source of our questions is manually translated logical reasoning questions from the Chinese Civil Service Examination. Experiments show that VisualPuzzles requires significantly less intensive domain-specific knowledge and more complex reasoning compared to benchmarks like MMMU, enabling us to better evaluate genuine multimodal reasoning. Evaluations show that state-of-the-art multimodal large language models consistently lag behind human performance on VisualPuzzles, and that strong performance on knowledge-intensive benchmarks does not necessarily translate to success on reasoning-focused, knowledge-light tasks. Additionally, reasoning enhancements such as scaling up inference compute (with "thinking" modes) yield inconsistent gains across models and task types, and we observe no clear correlation between model size and performance. We also found that models exhibit different reasoning and answering patterns on VisualPuzzles compared to benchmarks with heavier emphasis on knowledge. VisualPuzzles offers a clearer lens through which to evaluate reasoning capabilities beyond factual recall and domain knowledge.

The rise of LLM-driven AI characters raises safety concerns, particularly for vulnerable human users with psychological disorders. To address these risks, we propose EmoAgent, a multi-agent AI framework designed to evaluate and mitigate mental health hazards in human-AI interactions. EmoAgent comprises two components: EmoEval simulates virtual users, including those portraying mentally vulnerable individuals, to assess mental health changes before and after interactions with AI characters. It uses clinically proven psychological and psychiatric assessment tools (PHQ-9, PDI, PANSS) to evaluate mental risks induced by LLM. EmoGuard serves as an intermediary, monitoring users' mental status, predicting potential harm, and providing corrective feedback to mitigate risks. Experiments conducted in popular character-based chatbots show that emotionally engaging dialogues can lead to psychological deterioration in vulnerable users, with mental state deterioration in more than 34.4% of the simulations. EmoGuard significantly reduces these deterioration rates, underscoring its role in ensuring safer AI-human interactions. Our code is available at: https://github.com/1akaman/EmoAgent

A study that used artificial intelligence–generated content to "participate" in online discussions and test whether AI was more successful at changing people's minds than human-generated content has caused an uproar because of ethical concerns about the work. This week some of the unwitting research participants publicly asked the University of Zürich (UZH), where the researchers behind the experiment hold positions, to investigate and apologize. "I think people have a reasonable expectation to not be in scientific experiments without their consent," says Casey Fiesler, an expert on internet research ethics at the University of Colorado Boulder. A university statement emailed to Science says the researchers--who remain anonymous--have decided not to publish their results. The university will investigate the incident, the statement says.

Microsoft released its quarterly earnings report on Wednesday after the New York stock market closed, beating Wall Street's expectations for the fourth quarter in a row amid a financial boom for artificial intelligence businesses. The company revealed revenue of 70.07bn and earnings of 3.46 per share. The result exceeded analyst predictions that revenue would grow to 68.42bn, or 10.6% year-over-year, and that earnings-per-share would come in at 3.22. Shares in the tech giant jumped more than 5% in after-hours trading. Analysts viewed the earnings report as a temperature check on Microsoft's artificial intelligence (AI) business, which has announced it will invest around 80bn in this fiscal year alone, though it has also terminated some data center leases in recent months.