Analyzing and processing vast amounts of textual data presents significant challenges in efficiently extracting key information.In this paper, we introduce '***Spotlight***’, a novel paradigm for information extraction that produces concise, engaging narratives by highlighting the most compelling aspects of a document. Unlike highlights (fragmented key points) and traditional summaries, which prioritize comprehensive coverage, spotlights selectively emphasize intriguing content to foster deeper reader engagement with the source material. We formally differentiate spotlights from related constructs and support our analysis with a detailed benchmarking study using new datasets curated for this work. To generate high-quality spotlights, we propose a two-stage approach: fine-tuning a large language model on our benchmark data, followed by alignment via Direct Preference Optimization (DPO). Our comprehensive evaluation demonstrates that the resulting model not only identifies key elements with precision but also enhances readability and boosts the engagement value of the original document. Datasets and code are available at https://github.com/ankan2/Spotlight-EMNLP2025.

Digital tool-based agents, powered by Large Language Models (LLMs), that invoke external Application Programming Interfaces (APIs) often rely on documentation to understand API functionality. However, such documentation is frequently missing, outdated, privatized, or inconsistent—hindering the development of reliable, general-purpose agents. In this work, we propose a new research direction: learning of API functionality directly from in-context demonstrations. This task is a new paradigm applicable in scenarios without documentation. Using API benchmarks, we collect demonstrations from both expert agents and from self-exploration. To understand what information demonstrations must convey for successful task completion, we extensively study how the number of demonstrations and the use of LLM-generated summaries and evaluations affect the task success rate of the API-based agent. Our experiments across 3 datasets and 6 models show that learning functionality from in-context demonstrations remains a non-trivial challenge, even for state-of-the-art LLMs. We find that providing explicit function calls and natural language critiques significantly improves the agent’s task success rate due to more accurate parameter filling. We analyze failure modes, identify sources of error, and highlight key open challenges for future work in documentation-free, self-improving, API-based agents.

Visual content has been proven to be effective for micro-learning compared to other media. In this paper, we discuss leveraging this observation in our efforts to build audio-visual content for young learners’ vocabulary learning. We attempt to tackle two major issues in the process of traditional visual curation tasks. Generic learning videos do not necessarily satisfy the unique context of a learner and/or an educator, and hence may not result in maximal learning outcomes. Also, manual video curation by educators is a highly labor-intensive process. To this end, we present a customizable micro-learning audio-visual content curation tool that is designed to reduce the human (educator) effort in creating just-in-time learning videos from a textual description (learning script). This provides educators with control of the content while preparing the learning scripts, and in turn can also be customized to capture the desired learning objectives and outcomes. As a use case, we automatically generate learning videos with British National Corpus’ (BNC) frequently spoken vocabulary words and evaluate them with experts. They positively recommended the generated learning videos with an average rating of 4.25 on a Likert scale of 5 points. The inter-annotator agreement between the experts for the video quality was substantial (Fleiss Kappa=0.62) with an overall agreement of 81%.