Jennifer Healey

Peter L. Bartlett, Victor Gabillon, Jennifer Healey et al.

We address the problem of planning in an environment with deterministic dynamics and stochastic rewards with discounted returns. The optimal value function is not known, nor are the rewards bounded. We propose Platypoos, a simple scale-free planning algorithm that adapts to the unknown scale and smoothness of the reward function. We provide a sample complexity analysis for Platypoos that improves upon prior work and holds simultaneously over a broad range of discount factors and reward scales, without the algorithm knowing them. We also establish a matching lower bound showing our analysis is optimal up to constants.

Jinyan Su, Jennifer Healey

Pluralistic alignment requires systems to adapt to diverse user values, communication styles, and contextual assumptions. We believe that a foundational prerequisite for such alignment enabling accurate preference elicitation from people when their intent is under-specified or ambiguous. We study the problem of preference elicitation in multi-turn question answering by decomposing the problem into two components: a \textbf{clarification policy}, which decides whether to ask a clarifying question or answer directly, and \textbf{post-clarification answering}, which produces the correct final answer once the missing information is provided. We show, using the PACIFIC benchmark, that supervised fine-tuning rapidly improves the clarification policy, however, final answer accuracy remains substantially lower even when the model takes the correct action. This gap indicates that understanding and correctly interpreting the user's response is the critical gap in multi-turn question-answering systems.

Jinyan Su, Jennifer Healey, Preslav Nakov et al.

Retrieval-Augmented Generation (RAG) has emerged as a powerful approach to mitigate large language model (LLM) hallucinations by incorporating external knowledge retrieval. However, existing RAG frameworks often apply retrieval indiscriminately,leading to inefficiencies-over-retrieving when unnecessary or failing to retrieve iteratively when required for complex reasoning. Recent adaptive retrieval strategies, though adaptively navigates these retrieval strategies, predict only based on query complexity and lacks user-driven flexibility, making them infeasible for diverse user application needs. In this paper, we introduce a novel user-controllable RAG framework that enables dynamic adjustment of the accuracy-cost trade-off. Our approach leverages two classifiers: one trained to prioritize accuracy and another to prioritize retrieval efficiency. Via an interpretable control parameter $α$, users can seamlessly navigate between minimal-cost retrieval and high-accuracy retrieval based on their specific requirements. We empirically demonstrate that our approach effectively balances accuracy, retrieval cost, and user controllability, making it a practical and adaptable solution for real-world applications. Code is available at https://github.com/JinyanSu1/Flare-Aug.

Jinyan Su, Jennifer Healey, Preslav Nakov et al.

Large language models (LLMs) are increasingly optimized for long reasoning, under the assumption that more reasoning leads to better performance. However, emerging evidence suggests that longer responses can sometimes degrade accuracy rather than improve it. In this paper, we conduct a systematic empirical study of the relationship between reasoning length and answer correctness. We find that LLMs tend to overthink simple problems, generating unnecessarily long outputs, and underthink harder ones, failing to extend their reasoning when it is most needed. This indicates that models might misjudge problem difficulty and fail to calibrate their response length appropriately. Furthermore, we investigate the effects of length reduction with a preference optimization algorithm when simply preferring the shorter responses regardless of answer correctness. Experiments show that the generation length can be significantly reduced while maintaining acceptable accuracy. Our findings highlight generation length as a meaningful signal for reasoning behavior and motivate further exploration into LLMs' self-awareness in reasoning length adaptation.

Pierre Perrault, Jennifer Healey, Zheng Wen et al.

We introduce a new budgeted framework for online influence maximization, considering the total cost of an advertising campaign instead of the common cardinality constraint on a chosen influencer set. Our approach better models the real-world setting where the cost of influencers varies and advertisers want to find the best value for their overall social advertising budget. We propose an algorithm assuming an independent cascade diffusion model and edge level semi-bandit feedback, and provide both theoretical and experimental results. Our analysis is also valid for the cardinality constraint setting and improves the state of the art regret bound in this case.

Jennifer Healey, Laurie Byrum, Md Nadeem Akhtar et al.

Pre-trained large language models (LLMs) can now be easily adapted for specific business purposes using custom prompts or fine tuning. These customizations are often iteratively re-engineered to improve some aspect of performance, but after each change businesses want to ensure that there has been no negative impact on the system's behavior around such critical issues as bias. Prior methods of benchmarking bias use techniques such as word masking and multiple choice questions to assess bias at scale, but these do not capture all of the nuanced types of bias that can occur in free response answers, the types of answers typically generated by LLM systems. In this paper, we identify several kinds of nuanced bias in free text that cannot be similarly identified by multiple choice tests. We describe these as: confidence bias, implied bias, inclusion bias and erasure bias. We present a semi-automated pipeline for detecting these types of bias by first eliminating answers that can be automatically classified as unbiased and then co-evaluating name reversed pairs using crowd workers. We believe that the nuanced classifications our method generates can be used to give better feedback to LLMs, especially as LLM reasoning capabilities become more advanced.

Wanrong Zhu, Jennifer Healey, Ruiyi Zhang et al.

Recent advancements in instruction-following models have made user interactions with models more user-friendly and efficient, broadening their applicability. In graphic design, non-professional users often struggle to create visually appealing layouts due to limited skills and resources. In this work, we introduce a novel multimodal instruction-following framework for layout planning, allowing users to easily arrange visual elements into tailored layouts by specifying canvas size and design purpose, such as for book covers, posters, brochures, or menus. We developed three layout reasoning tasks to train the model in understanding and executing layout instructions. Experiments on two benchmarks show that our method not only simplifies the design process for non-professionals but also surpasses the performance of few-shot GPT-4V models, with mIoU higher by 12% on Crello. This progress highlights the potential of multimodal instruction-following models to automate and simplify the design process, providing an approachable solution for a wide range of design tasks on visually-rich documents.

Aishwarya Sahoo, Jeevana Kruthi Karnuthala, Tushar Parmanand Budhwani et al.

LLM-as-a-judge is a framework where a large language model (LLM) evaluates the output of another LLM. While LLMs excel at producing qualitative textual evaluations, they often struggle to predict human preferences and numeric scores. We propose quantitative LLM judges, which align evaluation scores of existing LLM judges to humans in a given domain using regression models. The models are trained to improve the score of the original judge using its rationale and score. We present four quantitative judges for different types of absolute and relative feedback, which showcases the generality and versatility of our framework. Our framework is more computationally efficient than supervised fine-tuning and can be more statistically efficient when human feedback is limited, which is expected in practice. We validate these claims empirically on four datasets using two base judges. Our experiments show that quantitative judges can improve the predictive power of existing judges through post-hoc modeling.

Vishakh Padmakumar, Zichao Wang, David Arbour et al.

While large language models (LLMs) are increasingly capable of handling longer contexts, recent work has demonstrated that they exhibit the "lost in the middle" phenomenon (Liu et al., 2024) of unevenly attending to different parts of the provided context. This hinders their ability to cover diverse source material in multi-document summarization, as noted in the DiverseSumm benchmark (Huang et al., 2024). In this work, we contend that principled content selection is a simple way to increase source coverage on this task. As opposed to prompting an LLM to perform the summarization in a single step, we explicitly divide the task into three steps -- (1) reducing document collections to atomic key points, (2) using determinantal point processes (DPP) to perform select key points that prioritize diverse content, and (3) rewriting to the final summary. By combining prompting steps, for extraction and rewriting, with principled techniques, for content selection, we consistently improve source coverage on the DiverseSumm benchmark across various LLMs. Finally, we also show that by incorporating relevance to a provided user intent into the DPP kernel, we can generate personalized summaries that cover relevant source information while retaining coverage.

Jennifer Healey, Laurie Byrum, Md Nadeem Akhtar et al.

LLM evaluation is challenging even the case of base models. In real world deployments, evaluation is further complicated by the interplay of task specific prompts and experiential context. At scale, bias evaluation is often based on short context, fixed choice benchmarks that can be rapidly evaluated, however, these can lose validity when the LLMs' deployed context differs. Large scale human evaluation is often seen as too intractable and costly. Here we present our journey towards developing a semi-automated bias evaluation framework for free text responses that has human insights at its core. We discuss how we developed an operational definition of bias that helped us automate our pipeline and a methodology for classifying bias beyond multiple choice. We additionally comment on how human evaluation helped us uncover problematic templates in a bias benchmark.

Victor S. Bursztyn, Jennifer Healey, Vishwa Vinay

Based on recent advances in realistic language modeling (GPT-3) and cross-modal representations (CLIP), Gaudí was developed to help designers search for inspirational images using natural language. In the early stages of the design process, with the goal of eliciting a client's preferred creative direction, designers will typically create thematic collections of inspirational images called "mood-boards". Creating a mood-board involves sequential image searches which are currently performed using keywords or images. Gaudí transforms this process into a conversation where the user is gradually detailing the mood-board's theme. This representation allows our AI to generate new search queries from scratch, straight from a project briefing, following a theme hypothesized by GPT-3. Compared to previous computational approaches to mood-board creation, to the best of our knowledge, ours is the first attempt to represent mood-boards as the stories that designers tell when presenting a creative direction to a client.

Sridhar Mahadevan, Anup Rao, Georgios Theocharous et al.

Many real-world applications require aligning two temporal sequences, including bioinformatics, handwriting recognition, activity recognition, and human-robot coordination. Dynamic Time Warping (DTW) is a popular alignment method, but can fail on high-dimensional real-world data where the dimensions of aligned sequences are often unequal. In this paper, we show that exploiting the multiscale manifold latent structure of real-world data can yield improved alignment. We introduce a novel framework called Warping on Wavelets (WOW) that integrates DTW with a a multi-scale manifold learning framework called Diffusion Wavelets. We present a theoretical analysis of the WOW family of algorithms and show that it outperforms previous state of the art methods, such as canonical time warping (CTW) and manifold warping, on several real-world datasets.

Victor S. Bursztyn, Jennifer Healey, Nedim Lipka et al.

Conversations aimed at determining good recommendations are iterative in nature. People often express their preferences in terms of a critique of the current recommendation (e.g., "It doesn't look good for a date"), requiring some degree of common sense for a preference to be inferred. In this work, we present a method for transforming a user critique into a positive preference (e.g., "I prefer more romantic") in order to retrieve reviews pertaining to potentially better recommendations (e.g., "Perfect for a romantic dinner"). We leverage a large neural language model (LM) in a few-shot setting to perform critique-to-preference transformation, and we test two methods for retrieving recommendations: one that matches embeddings, and another that fine-tunes an LM for the task. We instantiate this approach in the restaurant domain and evaluate it using a new dataset of restaurant critiques. In an ablation study, we show that utilizing critique-to-preference transformation improves recommendations, and that there are at least three general cases that explain this improved performance.

Sofie Beier, Sam Berlow, Esat Boucaud et al.

Readability is on the cusp of a revolution. Fixed text is becoming fluid as a proliferation of digital reading devices rewrite what a document can do. As past constraints make way for more flexible opportunities, there is great need to understand how reading formats can be tuned to the situation and the individual. We aim to provide a firm foundation for readability research, a comprehensive framework for modern, multi-disciplinary readability research. Readability refers to aspects of visual information design which impact information flow from the page to the reader. Readability can be enhanced by changes to the set of typographical characteristics of a text. These aspects can be modified on-demand, instantly improving the ease with which a reader can process and derive meaning from text. We call on a multi-disciplinary research community to take up these challenges to elevate reading outcomes and provide the tools to do so effectively.

Namrata Srivastava, Rajiv Jain, Jennifer Healey et al.

As reading on mobile devices is becoming more ubiquitous, content is consumed in shorter intervals and is punctuated by frequent interruptions. In this work, we explore the best way to mitigate the effects of reading interruptions on longer text passages. Our hypothesis is that short summaries of either previously read content (reviews) or upcoming content (previews) will help the reader re-engage with the reading task. Our target use case is for students who study using electronic textbooks and who are frequently mobile. We present a series of pilot studies that examine the benefits of different types of summaries and their locations, with respect to variations in text content and participant cohorts. We find that users prefer reviews after an interruption, but that previews shown after interruptions have a larger positive influence on comprehension. Our work is a first step towards smart reading applications that proactively provide text summaries to mitigate interruptions on the go.

Victor S. Bursztyn, Jennifer Healey, Eunyee Koh et al.

We have developed a conversational recommendation system designed to help users navigate through a set of limited options to find the best choice. Unlike many internet scale systems that use a singular set of search terms and return a ranked list of options from amongst thousands, our system uses multi-turn user dialog to deeply understand the users preferences. The system responds in context to the users specific and immediate feedback to make sequential recommendations. We envision our system would be highly useful in situations with intrinsic constraints, such as finding the right restaurant within walking distance or the right retail item within a limited inventory. Our research prototype instantiates the former use case, leveraging real data from Google Places, Yelp, and Zomato. We evaluated our system against a similar system that did not incorporate user feedback in a 16 person remote study, generating 64 scenario-based search journeys. When our recommendation system was successfully triggered, we saw both an increase in efficiency and a higher confidence rating with respect to final user choice. We also found that users preferred our system (75%) compared with the baseline.

Duotun Wang, Jennifer Healey, Jing Qian et al.

We present the result of a pilot study measuring child engagement with the Lets Make A Story system, a novel mixed reality, MR, collaborative storytelling system designed for grandparents and grandchildren. We compare our MR experience against an equivalent paper story experience. The goal of our pilot was to test the system with actual child users and assess the goodness of using metrics of time, user generated story content and facial expression analysis as metrics of child engagement. We find that multiple confounding variables make these metrics problematic including attribution of engagement time, spontaneous non-story related conversation and having the childs full forward face continuously in view during the story. We present our platform and experiences and our finding that the strongest metric was user comments in the post-experiential interview.