Large Language Models (LLMs) require sophisticated prompting, yet current practices face challenges in structure, data integration, format sensitivity, and tooling. Existing methods lack comprehensive solutions for organizing complex prompts involving diverse data types (documents, tables, images) or managing presentation variations systematically. To address these gaps, we introduce POML (Prompt Orchestration Markup Language). POML employs component-based markup for logical structure (roles, tasks, examples), specialized tags for seamless data integration, and a CSS-like styling system to decouple content from presentation, reducing formatting sensitivity. It includes templating for dynamic prompts and a comprehensive developer toolkit (IDE support, SDKs) to improve version control and collaboration. We validate POML through two case studies demonstrating its impact on complex application integration (PomLink) and accuracy performance (TableQA), as well as a user study assessing its effectiveness in real-world development scenarios.

Objective Structured Clinical Examinations (OSCEs) are essential for medical training, but they require significant resources, including professional actors and expert medical feedback. Although Large Language Models (LLMs) have introduced text-based virtual patients for communication practice, these simulations often lack the capability for richer, non-textual interactions. This paper presents a novel framework that significantly enhances LLM-based simulated patients by equipping them with action spaces, thereby enabling more realistic and dynamic patient behaviors that extend beyond text. Furthermore, our system incorporates virtual tutors that provide students with instant, personalized feedback on their performance at any time during these simulated encounters. We have conducted a rigorous evaluation of the framework's real-time performance, including system latency and component accuracy. Preliminary evaluations with medical experts assessed the naturalness and coherence of the simulated patients, as well as the usefulness and appropriateness of the virtual tutor's assessments. This innovative system provides medical students with a low-cost, accessible platform for personalized OSCE preparation at home.

Recently, multimodal large language models (MLLMs) have achieved significant advancements across various domains, and corresponding evaluation benchmarks have been continuously refined and improved. In this process, benchmarks in the scientific domain have played an important role in assessing the reasoning capabilities of MLLMs. However, existing benchmarks still face three key challenges: 1) Insufficient evaluation of models' reasoning abilities in multilingual scenarios; 2) Inadequate assessment of MLLMs' comprehensive modality coverage; 3) Lack of fine-grained annotation of scientific knowledge points. To address these gaps, we propose MME-SCI, a comprehensive and challenging benchmark. We carefully collected 1,019 high-quality question-answer pairs, which involve 3 distinct evaluation modes. These pairs cover four subjects, namely mathematics, physics, chemistry, and biology, and support five languages: Chinese, English, French, Spanish, and Japanese. We conducted extensive experiments on 16 open-source models and 4 closed-source models, and the results demonstrate that MME-SCI is widely challenging for existing MLLMs. For instance, under the Image-only evaluation mode, o4-mini achieved accuracy of only 52.11%, 24.73%, 36.57%, and 29.80% in mathematics, physics, chemistry, and biology, respectively, indicating a significantly higher difficulty level compared to existing benchmarks. More importantly, using MME-SCI's multilingual and fine-grained knowledge attributes, we analyzed existing models' performance in depth and identified their weaknesses in specific domains. The Data and Evaluation Code are available at https://github.com/JCruan519/MME-SCI.

This work revisits and extends synthetic query generation pipelines for Neural Information Retrieval (NIR) by leveraging the InPars Toolkit, a reproducible, end-to-end framework for generating training data using large language models (LLMs). We first assess the reproducibility of the original InPars, InPars-V2, and Promptagator pipelines on the SciFact benchmark and validate their effectiveness using open-source reranker and generator models. Building on this foundation, we introduce two key extensions to the pipeline: (1) fine-tuning a query generator LLM via Contrastive Preference Optimization (CPO) to improve the signal quality in generated queries, and (2) replacing static prompt templates with dynamic, Chain-of-Thought (CoT) optimized prompts using the DSPy framework. Our results show that both extensions reduce the need for aggressive filtering while improving retrieval performance. All code, models, and synthetic datasets are publicly released to support further research at: \href{https://github.com/danilotpnta/IR2-project}{this https URL}.

Embodied AI (EAI) agents continuously interact with the physical world, generating vast, heterogeneous multimodal data streams that traditional management systems are ill-equipped to handle. In this survey, we first systematically evaluate five storage architectures (Graph Databases, Multi-Model Databases, Data Lakes, Vector Databases, and Time-Series Databases), focusing on their suitability for addressing EAI's core requirements, including physical grounding, low-latency access, and dynamic scalability. We then analyze five retrieval paradigms (Fusion Strategy-Based Retrieval, Representation Alignment-Based Retrieval, Graph-Structure-Based Retrieval, Generation Model-Based Retrieval, and Efficient Retrieval-Based Optimization), revealing a fundamental tension between achieving long-term semantic coherence and maintaining real-time responsiveness. Based on this comprehensive analysis, we identify key bottlenecks, spanning from the foundational Physical Grounding Gap to systemic challenges in cross-modal integration, dynamic adaptation, and open-world generalization. Finally, we outline a forward-looking research agenda encompassing physics-aware data models, adaptive storage-retrieval co-optimization, and standardized benchmarking, to guide future research toward principled data management solutions for EAI. Our survey is based on a comprehensive review of more than 180 related studies, providing a rigorous roadmap for designing the robust, high-performance data management frameworks essential for the next generation of autonomous embodied systems.

--Existing driving style recognition systems largely depend on low-level sensor-derived features for training, neglecting the rich semantic reasoning capability inherent to human experts. T o bridge this gap, we propose a novel framework that integrates Semantic Privileged Information (SPI) derived from large language models (LLMs) to align recognition outcomes with human-interpretable reasoning. First, we introduce DriBehavGPT, an interactive LLM-based module that generates natural-language descriptions of driving behaviors. These descriptions are then encoded into machine learning-compatible representations via text embedding and dimensionality reduction. Finally, we incorporate them as privileged information into Support V ector Machine Plus (SVM+) for training, enabling the model to approximate human-like interpretation patterns. Importantly, SPI is exclusively used during training, while inference relies solely on sensor data, ensuring computational efficiency without sacrificing performance. COGNIZING driving styles plays a pivotal role in understanding human-vehicle interactions, thereby improving personalized driving experience and enhancing the acceptance of advanced driver assistance systems [1]. For example, adaptive cruise control systems offer configurable parameters, such as inter-vehicle distance, target speed, and driving modes, to accommodate both aggressive drivers prioritizing traffic throughput efficiency and conservative drivers emphasizing safety [2], [3].

Reinforcement learning (RL) has demonstrated great potential in robotic operations. However, its data-intensive nature and reliance on the Markov Decision Process (MDP) assumption limit its practical deployment in real-world scenarios involving complex dynamics and long-term temporal dependencies, such as multi-robot manipulation. Decision Transformers (DTs) have emerged as a promising offline alternative by leveraging causal transformers for sequence modeling in RL tasks. However, their applications to multi-robot manipulations still remain underexplored. To address this gap, we propose a novel framework, Symbolically-Guided Decision Transformer (SGDT), which integrates a neuro-symbolic mechanism with a causal transformer to enable deployable multi-robot collaboration. In the proposed SGDT framework, a neuro-symbolic planner generates a high-level task-oriented plan composed of symbolic subgoals. Guided by these subgoals, a goal-conditioned decision transformer (GCDT) performs low-level sequential decision-making for multi-robot manipulation. This hierarchical architecture enables structured, interpretable, and generalizable decision making in complex multi-robot collaboration tasks. We evaluate the performance of SGDT across a range of task scenarios, including zero-shot and few-shot scenarios. To our knowledge, this is the first work to explore DT-based technology for multi-robot manipulation.

LLMs have recently been used to generate Python programs representing generalized plans in PDDL planning, i.e., plans that generalize across the tasks of a given PDDL domain. Previous work proposed a framework consisting of three steps: the LLM first generates a summary and then a strategy for the domain, both in natural language, and then implements that strategy as a Python program, that gets debugged on example planning tasks. In that work, only one strategy is generated and passed directly to the program generation. If the strategy is incorrect, its implementation will therefore result in an incorrect generalized plan. Here, we introduce an approach that generates the strategy in the form of pseudocode and enables automatic debugging of the pseudocode, hence allowing us to identify and fix errors prior to the generation of the generalized plan itself. Additionally, we extend the Python debugging phase with a reflection step prompting the LLM to pinpoint the reason for the observed plan failure. Finally, we take inspiration from LLM code generation to produce several program variants and pick the best one. Running experiments on 17 benchmark domains, we show that these extensions substantially improve (and never deteriorate) the quality of the generalized plans. In 12 of the domains, our best Python programs solve all tasks that can be generated with the respective instance generator.

Evaluating Natural Language Generation (NLG) is crucial for the practical adoption of AI, but has been a longstanding research challenge. While human evaluation is considered the de-facto standard, it is expensive and lacks scalability. Practical applications have driven the development of various automatic evaluation metrics (AEM), designed to compare the model output with human-written references, generating a score which approximates human judgment. Over time, AEMs have evolved from simple lexical comparisons, to semantic similarity models and, more recently, to LLM-based evaluators. However, it seems that no single metric has emerged as a definitive solution, resulting in studies using different ones without fully considering the implications. This paper aims to show this by conducting a thorough examination of the methodologies of existing metrics, their documented strengths and limitations, validation methods, and correlations with human judgment. We identify several key challenges: metrics often capture only specific aspects of text quality, their effectiveness varies by task and dataset, validation practices remain unstructured, and correlations with human judgment are inconsistent. Importantly, we find that these challenges persist in the most recent type of metric, LLM-as-a-Judge, as well as in the evaluation of Retrieval Augmented Generation (RAG), an increasingly relevant task in academia and industry. Our findings challenge the quest for the 'perfect metric'. We propose selecting metrics based on task-specific needs and leveraging complementary evaluations and advocate that new metrics should focus on enhanced validation methodologies.

Controlling the length of text produced by large language models (LLMs) remains challenging: models frequently overshoot or undershoot explicit length instructions because they cannot reliably keep an internal token count. We present a prompt-based, one-shot strategy that compels an off-the-shelf LLM to generate exactly a desired number of tokens--words (English) or characters (Chinese)--without any fine-tuning or iterative sampling. The prompt appends countdown markers and explicit counting rules so that the model "writes while counting." We evaluate on four settings: open-ended generation (1-1000 tokens), XSUM summarization, MT-Bench-LI instruction following, and the LIFEBENCH equal-length track. On MT-Bench-LI, strict length compliance with gpt-4.1 leaps from below 30 % under na ıve prompts to above 95 % with our countdown prompt--surpassing the popular draft-then-revise baseline--while judged answer quality is preserved. These results demonstrate that precise length control can be achieved through prompt engineering alone, offering a lightweight alternative to training-or decoding-based methods.