Nishant Balepur, Malachi Hamada, Varsha Kishore, Sergey Feldman, Amanpreet Singh, Pao Siangliulue, Joseph Chee Chang, Eunsol Choi, Jordan Lee Boyd-Graber, Aakanksha Naik. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

As AI agents take on increasingly long-running tasks involving sophisticated planning and execution, there is a corresponding need for novel interaction designs that enable deeper human-agent collaboration. However, most prior works leverage human interaction to fix “autonomous” workflows that have yet to become fully autonomous or rigidly treat planning and execution as separate stages. Based on a formative study with 9 researchers using AI to support their work, we propose a design that affords greater flexibility in collaboration, so that users can 1) delegate agency to the user or agent via a collaborative plan where individual steps can be assigned; and 2) interleave planning and execution so that plans can adjust after partial execution. We introduce Cocoa, a system that takes design inspiration from computational notebooks to support complex research tasks. A lab study (n = 16) found that Cocoa enabled steerability without sacrificing ease-of-use, and a week-long field deployment (n = 7) showed how researchers collaborated with Cocoa to accomplish real-world tasks.

We present SciArena, an open and collaborative platform for evaluating foundation models on scientific literature-grounded tasks. Unlike traditional benchmarks for scientific literature understanding and synthesis, SciArena engages the research community directly, following the Chatbot Arena evaluation approach of community voting on model comparisons.By leveraging collective intelligence, SciArena offers a community-driven evaluation of model performance on open-ended scientific tasks that demand literature-grounded, long-form responses.The platform currently supports 44 open-source and proprietary foundation models and has collected over 19,000 votes from human researchers across diverse scientific domains. Our analysis of the data collected so far confirms its high quality.We discuss the results and insights based on the model ranking leaderboard.To further promote research in building model-based automated evaluation systems for literature tasks, we release SciArena-Eval, a meta-evaluation benchmark based on our collected preference data. The benchmark measures the accuracy of models in judging answer quality by comparing their pairwise assessments with human votes. Our experiments highlight the benchmark’s challenges and emphasize the need for more reliable automated evaluation methods.

Abstract Language model users often issue queries that lack specification, where the context under which a query was issued—such as the user’s identity, the query’s intent, and the criteria for a response to be useful—is not explicit. For instance, a good response to a subjective query like “What book should I read next?” would depend on the user’s preferences, and a good response to an open-ended query like “How do antibiotics work against bacteria?” would depend on the user’s expertise. This makes evaluation of responses to such queries an ill-posed task, as evaluators may make arbitrary judgments about the response quality. To remedy this, we present contextualized evaluations, a protocol that synthetically constructs context surrounding an underspecified query and provides it during evaluation. We find that the presence of context can 1) alter conclusions drawn from evaluation, even flipping benchmark rankings between model pairs, 2) nudge evaluators to make fewer judgments based on surface-level criteria, like style, and 3) provide new insights about model behavior across diverse contexts. Specifically, our procedure suggests a potential bias towards WEIRD (Western, Educated, Industrialized, Rich and Democratic) contexts in models’ “default” responses and we find that models are not equally sensitive to following different contexts, even when they are provided in prompts.1

In our era of rapid technological advancement, the research landscape for writing assistants has become increasingly fragmented across various research communities. We seek to address this challenge by proposing a design space as a structured way to examine and explore the multidimensional space of intelligent and interactive writing assistants. Through community collaboration, we explore five aspects of writing assistants: task, user, technology, interaction, and ecosystem. Within each aspect, we define dimensions and codes by systematically reviewing 115 papers, while leveraging the expertise of researchers in various disciplines. Our design space aims to offer researchers and designers a practical tool to navigate, comprehend, and compare the various possibilities of writing assistants, and aid in the design of new writing assistants.

With the rapid growth of scholarly archives, researchers subscribe to “paper alert’’ systems that periodically provide them with recommendations of recently published papers that are similar to previously collected papers. However, researchers sometimes struggle to make sense of nuanced connections between recommended papers and their own research context, as existing systems only present paper titles and abstracts. To help researchers spot these connections, we present PaperWeaver, an enriched paper alerts system that provides contextualized text descriptions of recommended papers based on user-collected papers. PaperWeaver employs a computational method based on Large Language Models (LLMs) to infer users’ research interests from their collected papers, extract context-specific aspects of papers, and compare recommended and collected papers on these aspects. Our user study (N=15) showed that participants using PaperWeaver were able to better understand the relevance of recommended papers and triage them more confidently when compared to a baseline that presented the related work sections from recommended papers.

In communities with social hierarchies, fear of judgment can discourage communication. While anonymity may alleviate some social pressure, fully anonymous spaces enable toxic behavior and hide the social context that motivates people to participate and helps them tailor their communication. We explore a design space of meronymous communication, where people can reveal carefully chosen aspects of their identity and also leverage trusted endorsers to gain credibility. We implemented these ideas in a system for scholars to meronymously seek and receive paper recommendations on Twitter and Mastodon. A formative study with 20 scholars confirmed that scholars see benefits to participating but are deterred due to social anxiety. From a month-long public deployment, we found that with meronymity, junior scholars could comfortably ask “newbie” questions and get responses from senior scholars who they normally found intimidating. Responses were also tailored to the aspects about themselves that junior scholars chose to reveal.

Benjamin Newman, Yoonjoo Lee, Aakanksha Naik, Pao Siangliulue, Raymond Fok, Juho Kim, Daniel S Weld, Joseph Chee Chang, Kyle Lo. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

Scientific jargon can confuse researchers when they read materials from other domains. Identifying and translating jargon for individual researchers could speed up research, but current methods of jargon identification mainly use corpus-level familiarity indicators rather than modeling researcher-specific needs, which can vary greatly based on each researcher's background. We collect a dataset of over 10K term familiarity annotations from 11 computer science researchers for terms drawn from 100 paper abstracts. Analysis of this data reveals that jargon familiarity and information needs vary widely across annotators, even within the same sub-domain (e.g., NLP). We investigate features representing domain, subdomain, and individual knowledge to predict individual jargon familiarity. We compare supervised and prompt-based approaches, finding that prompt-based methods using information about the individual researcher (e.g., personal publications, self-defined subfield of research) yield the highest accuracy, though the task remains difficult and supervised approaches have lower false positive rates. This research offers insights into features and methods for the novel task of integrating personal data into scientific jargon identification.

Navigating the vast scientific literature often starts with browsing a paper’s abstract. However, when a reader seeks additional information, not present in the abstract, they face a costly cognitive chasm during their dive into the full text. To bridge this gap, we introduce recursively expandable abstracts, a novel interaction paradigm that dynamically expands abstracts by progressively incorporating additional information from the papers’ full text. This lightweight interaction allows scholars to specify their information needs by quickly brushing over the abstract or selecting AI-suggested expandable entities. Relevant information is synthesized using a retrieval-augmented generation approach, presented as a fluid, threaded expansion of the abstract, and made efficiently verifiable via attribution to relevant source-passages in the paper. Through a series of user studies, we demonstrate the utility of recursively expandable abstracts and identify future opportunities to support low-effort and just-in-time exploration of long-form information contexts through LLM-powered interactions.

When reading a scholarly article, inline citations help researchers contextualize the current article and discover relevant prior work. However, it can be challenging to prioritize and make sense of the hundreds of citations encountered during literature reviews. This paper introduces CiteSee, a paper reading tool that leverages a user’s publishing, reading, and saving activities to provide personalized visual augmentations and context around citations. First, CiteSee connects the current paper to familiar contexts by surfacing known citations a user had cited or opened. Second, CiteSee helps users prioritize their exploration by highlighting relevant but unknown citations based on saving and reading history. We conducted a lab study that suggests CiteSee is significantly more effective for paper discovery than three baselines. A field deployment study shows CiteSee helps participants keep track of their explorations and leads to better situational awareness and increased paper discovery via inline citation when conducting real-world literature reviews.

In order to help scholars understand and follow a research topic, significant research has been devoted to creating systems that help scholars discover relevant papers and authors. Recent approaches have shown the usefulness of highlighting relevant authors while scholars engage in paper discovery. However, these systems do not capture and utilize users’ evolving knowledge of authors. We reflect on the design space and introduce ComLittee, a literature discovery system that supports author-centric exploration. In contrast to paper-centric interaction in prior systems, ComLittee’s author-centric interaction supports curating research threads from individual authors, finding new authors and papers using combined signals from a paper recommender and the curated authors’ authorship graphs, and understanding them in the context of those signals. In a within-subjects experiment that compares to a paper-centric discovery system with author-highlighting, we demonstrate how ComLittee improves author and paper discovery.

Scholars who want to research a scientific topic must take time to read, extract meaning, and identify connections across many papers. As scientific literature grows, this becomes increasingly challenging. Meanwhile, authors summarize prior research in papers’ related work sections, though this is scoped to support a single paper. A formative study found that while reading multiple related work paragraphs helps overview a topic, it is hard to navigate overlapping and diverging references and research foci. In this work, we design a system, Relatedly, that scaffolds exploring and reading multiple related work paragraphs on a topic, with features including dynamic re-ranking and highlighting to spotlight unexplored dissimilar information, auto-generated descriptive paragraph headings, and low-lighting of redundant information. From a within-subjects user study (n=15), we found that scholars generate more coherent, insightful, and comprehensive topic outlines using Relatedly compared to a baseline paper list.

Research consumption has been traditionally limited to the reading of academic papers—a static, dense, and formally written format. Alternatively, pre-recorded conference presentation videos, which are more dynamic, concise, and colloquial, have recently become more widely available but potentially under-utilized. In this work, we explore the design space and benefits for combining academic papers and talk videos to leverage their complementary nature to provide a rich and fluid research consumption experience. Based on formative and co-design studies, we present Papeos, a novel reading and authoring interface that allow authors to augment their papers by segmenting and localizing talk videos alongside relevant paper passages with automatically generated suggestions. With Papeos, readers can visually skim a paper through clip thumbnails, and fluidly switch between consuming dense text in the paper or visual summaries in the video. In a comparative lab study (n=16), Papeos reduced mental load, scaffolded navigation, and facilitated more comprehensive reading of papers.

Efficiently reviewing scholarly literature and synthesizing prior art are crucial for scientific progress. Yet, the growing scale of publications and the burden of knowledge make synthesis of research threads more challenging than ever.While significant research has been devoted to helping scholars interact with individual papers, building research threads scattered across multiple papers remains a challenge.Most top-down synthesis (and LLMs) make it difficult to personalize and iterate on the output, while bottom-up synthesis is costly in time and effort.Here, we explore a new design space of mixed-initiative workflows.In doing so we develop a novel computational pipeline, Synergi, that ties together user input of relevant seed threads with citation graphs and LLMs, to expand and structure them, respectively.Synergiallows scholars to start with an entire threads-and-subthreads structure generated from papers relevant to their interests, and to iterate and customize on it as they wish. In our evaluation, we find that Synergi helps scholars efficiently make sense of relevant threads, broaden their perspectives, and increases their curiosity. We discuss future design implications for thread-based, mixed-initiative scholarly synthesis support tools.

Reviewing the literature to understand relevant threads of past work is a critical part of research and vehicle for learning. However, as the scientific literature grows the challenges for users to find and make sense of the many different threads of research grow as well. Previous work has helped scholars to find and group papers with citation information or textual similarity using standalone tools or overview visualizations. Instead, in this work we explore a tool integrated into users’ reading process that helps them with leveraging authors’ existing summarization of threads, typically in introduction or related work sections, in order to situate their own work’s contributions. To explore this we developed a prototype that supports efficient extraction and organization of threads along with supporting evidence as scientists read research articles. The system then recommends further relevant articles based on user-created threads. We evaluate the system in a lab study and find that it helps scientists to follow and curate research threads without breaking out of their flow of reading, collect relevant papers and clips, and discover interesting new articles to further grow threads.

People spend a significant amount of time trying to make sense of the internet, collecting content from a variety of sources and organizing it to make decisions and achieve their goals. While humans are able to fluidly iterate on collecting and organizing information in their minds, existing tools and approaches introduce significant friction into the process. We introduce Fuse, a browser extension that externalizes users’ working memory by combining low-cost collection with lightweight organization of content in a compact card-based sidebar that is always available. Fuse helps users simultaneously extract key web content and structure it in a lightweight and visual way. We discuss how these affordances help users externalize more of their mental model into the system (e.g., saving, annotating, and structuring items) and support fast reviewing and resumption of task contexts. Our 22-month public deployment and follow-up interviews provide longitudinal insights into the structuring behaviors of real-world users conducting information foraging tasks.

Consumers conducting comparison shopping, researchers making sense of competitive space, and developers looking for code snippets online all face the challenge of capturing the information they find for later use without interrupting their current flow. In addition, during many learning and exploration tasks, people need to externalize their mental context, such as estimating how urgent a topic is to follow up on, or rating a piece of evidence as a “pro” or “con,” which helps scaffold subsequent deeper exploration. However, current approaches incur a high cost, often requiring users to select, copy, context switch, paste, and annotate information in a separate document without offering specific affordances that capture their mental context. In this work, we explore a new interaction technique called “wiggling,” which can be used to fluidly collect, organize, and rate information during early sensemaking stages with a single gesture. Wiggling involves rapid back-and-forth movements of a pointer or up-and-down scrolling on a smartphone, which can indicate the information to be collected and its valence, using a single, light-weight gesture that does not interfere with other interactions that are already available. Through implementation and user evaluation, we found that wiggling helped participants accurately collect information and encode their mental context with a 58% reduction in operational cost while being 24% faster compared to a common baseline.

Tabs have become integral to browsing the Web yet have changed little since their introduction nearly 20 years ago. In contrast, the internet has gone through dramatic changes, with users increasingly moving from navigating to websites to exploring information across many sources to support online sensemaking. This paper investigates how tabs today are overloaded with a diverse set of functionalities and issues users face when managing them. We interviewed ten information workers asking about their tab management strategies and walk through each open tab on their work computers four times over two weeks. We uncovered competing pressures pushing for keeping tabs open (ranging from interaction to emotional costs) versus pushing for closing them (such as limited attention and resources). We then surveyed 103 participants to estimate the frequencies of these pressures at scale. Finally, we developed design implications for future browser interfaces that can better support managing these pressures.

Despite the increasing complexity and scale of people’s online activities, browser interfaces have stayed largely the same since tabs were introduced in major browsers nearly 20 years ago. The gap between simple tab-based browser interfaces and the complexity of users’ tasks can lead to serious adverse effects – commonly referred to as “tab overload.” This paper introduces a Chrome extension called Tabs.do, which explores bringing a task-centric approach to the browser, helping users to group their tabs into tasks and then organize, prioritize, and switch between those tasks fluidly. To lower the cost of importing, Tabs.do uses machine learning to make intelligent suggestions for grouping users’ open tabs into task bundles by exploiting behavioral and semantic features. We conducted a field deployment study where participants used Tabs.do with their real-life tasks in the wild, and showed that Tabs.do can decrease tab clutter, enabled users to create rich task structures with lightweight interactions, and allowed participants to context-switch among tasks more efficiently.

While there is an enormous amount of information online for making decisions such as choosing a product, restaurant, or school, it can be costly for users to synthesize that information into confident decisions. Information for users' many different criteria needs to be gathered from many different sources into a structure where they can be compared and contrasted. The usefulness of each criterion for differentiating potential options can be opaque to users, and evidence such as reviews may be subjective and conflicting, requiring users to interpret each under their personal context. We introduce Mesh, which scaffolds users to iteratively build up a better understanding of both their criteria and options by evaluating evidence gathered across sources in the context of consumer decision-making. Mesh bridges the gap between decision support systems that typically have rigid structures and the fluid and dynamic process of exploratory search, changing the cost structure to provide increasing payoffs with greater user investment. Our lab and field deployment studies found evidence that Mesh significantly reduces the costs of gathering and evaluating evidence and scaffolds decision-making through personalized criteria enabling users to gain deeper insights from data.

Whether figuring out where to eat in an unfamiliar city or deciding which apartment to live in, consumer generated data (i.e. reviews and forum posts) are often an important influence in online decision making. To make sense of these rich repositories of diverse opinions, searchers need to sift through a large number of reviews to characterize each item based on aspects that they care about. We introduce a novel system, SearchLens, where searchers build up a collection of "Lenses" that reflect their different latent interests, and compose the Lenses to find relevant items across different contexts. Based on the Lenses, SearchLens generates personalized interfaces with visual explanations that promotes transparency and enables deeper exploration. While prior work found searchers may not wish to put in effort specifying their goals without immediate and sufficient benefits, results from a controlled lab study suggest that our approach incentivized participants to express their interests more richly than in a baseline condition, and a field study showed that participants found benefits in SearchLens while conducting their own tasks.

People engaged in complex searches such as planning a vacation or understanding their medical symptoms are often overwhelmed by opening and managing many tabs. These challenges are exacerbated as search moves to smartphones and mobile devices where screen real-estate is limited and tasks are frequently suspended, resumed, and interleaved. Rather than continue to utilize tab-based browsing for complex search, we introduce a new way of browsing through a scaffolded interface. The list of search results serves as a mutable workspace, where a user can track progress on a specific information query. The search query serves as a gateway into this workspace, accessed through a task-subtask hierarchy. We instantiate this in the Bento mobile search system and investigate its effectiveness in three studies. We find converging evidence that users were able to make progress on their complex searching tasks with this structure, and find it more organized and easier to revisit.

Crowd-powered conversational assistants have been shown to be more robust than automated systems, but do so at the cost of higher response latency and monetary costs. A promising direction is to combine the two approaches for high quality, low latency, and low cost solutions. In this paper, we introduce Evorus, a crowd-powered conversational assistant built to automate itself over time by (i) allowing new chatbots to be easily integrated to automate more scenarios, (ii) reusing prior crowd answers, and (iii) learning to automatically approve response candidates. Our 5-month-long deployment with 80 participants and 281 conversations shows that Evorus can automate itself without compromising conversation quality. Crowd-AI architectures have long been proposed as a way to reduce cost and latency for crowd-powered systems; Evorus demonstrates how automation can be introduced successfully in a deployed system. Its architecture allows future researchers to make further innovation on the underlying automated components in the context of a deployed open domain dialog system.

Scientific discoveries are often driven by finding analogies in distant domains, but the growing number of papers makes it difficult to find relevant ideas in a single discipline, let alone distant analogies in other domains. To provide computational support for finding analogies across domains, we introduce SOLVENT, a mixed-initiative system where humans annotate aspects of research papers that denote their background (the high-level problems being addressed), purpose (the specific problems being addressed), mechanism (how they achieved their purpose), and findings (what they learned/achieved), and a computational model constructs a semantic representation from these annotations that can be used to find analogies among the research papers. We demonstrate that this system finds more analogies than baseline information-retrieval approaches; that annotators and annotations can generalize beyond domain; and that the resulting analogies found are useful to experts. These results demonstrate a novel path towards computationally supported knowledge sharing in research communities.

Crowdsourcing provides a scalable and efficient way to construct labeled datasets for training machine learning systems. However, creating comprehensive label guidelines for crowdworkers is often prohibitive even for seemingly simple concepts. Incomplete or ambiguous label guidelines can then result in differing interpretations of concepts and inconsistent labels. Existing approaches for improving label quality, such as worker screening or detection of poor work, are ineffective for this problem and can lead to rejection of honest work and a missed opportunity to capture rich interpretations about data. We introduce Revolt, a collaborative approach that brings ideas from expert annotation workflows to crowd-based labeling. Revolt eliminates the burden of creating detailed label guidelines by harnessing crowd disagreements to identify ambiguous concepts and create rich structures (groups of semantically related items) for post-hoc label decisions. Experiments comparing Revolt to traditional crowdsourced labeling show that Revolt produces high quality labels without requiring label guidelines in turn for an increase in monetary cost. This up front cost, however, is mitigated by Revolt's ability to produce reusable structures that can accommodate a variety of label boundaries without requiring new data to be collected. Further comparisons of Revolt's collaborative and non-collaborative variants show that collaboration reaches higher label accuracy with lower monetary cost.

Crowdsourced clustering approaches present a promising way to harness deep semantic knowledge for clustering complex information. However, existing approaches have difficulties supporting the global context needed for workers to generate meaningful categories, and are costly because all items require human judgments. We introduce Alloy, a hybrid approach that combines the richness of human judgments with the power of machine algorithms. Alloy supports greater global context through a new "sample and search" crowd pattern which changes the crowd's task from classifying a fixed subset of items to actively sampling and querying the entire dataset. It also improves efficiency through a two phase process in which crowds provide examples to help a machine cluster the head of the distribution, then classify low-confidence examples in the tail. To accomplish this, Alloy introduces a modular "cast and gather" approach which leverages a machine learning backbone to stitch together different types of judgment tasks.

Crowdsourcing offers a powerful new paradigm for online work. However, real world tasks are often interdependent, requiring a big picture view of the difference pieces involved. Existing crowdsourcing approaches that support such tasks -- ranging from Wikipedia to flash teams -- are bottlenecked by relying on a small number of individuals to maintain the big picture. In this paper, we explore the idea that a computational system can scaffold an emerging interdependent, big picture view entirely through the small contributions of individuals, each of whom sees only a part of the whole. To investigate the viability, strengths, and weaknesses of this approach we instantiate the idea in a prototype system for accomplishing distributed information synthesis and evaluate its output across a variety of topics. We also contribute a set of design patterns that may be informative for other systems aimed at supporting big picture thinking in small pieces.

Patients researching medical diagnoses, scientist exploring new fields of literature, and students learning about new domains are all faced with the challenge of capturing information they find for later use. However, saving information is challenging on mobile devices, where the small screen and font sizes combined with the inaccuracy of finger based touch screens makes it time consuming and stressful for people to select and save text for future use. Furthermore, beyond the challenge of simply selecting a region of bounded text on a mobile device, in many learning and data exploration tasks the boundaries of what text may be relevant and useful later are themselves uncertain for the user. In contrast to previous approaches which focused on speeding up the selection process by making the identification of hard boundaries faster, we introduce the idea of intentionally supporting uncertain input in the context of saving information during complex reading and information exploration. We embody this idea in a system that uses force touch and fuzzy bounding boxes along with posthoc expandable context to support identifying and saving information in an intentionally uncertain way on mobile devices. In a two part user study we find that this approach reduced selection time and was preferred by participants over the default system text selection method.