Yu Hou

Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao et al. · allen-ai, amazon-science

Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabilities are as yet poorly characterized. In order to inform future research, prepare for disruptive new model capabilities, and ameliorate socially harmful effects, it is vital that we understand the present and near-future capabilities and limitations of language models. To address this challenge, we introduce the Beyond the Imitation Game benchmark (BIG-bench). BIG-bench currently consists of 204 tasks, contributed by 450 authors across 132 institutions. Task topics are diverse, drawing problems from linguistics, childhood development, math, common-sense reasoning, biology, physics, social bias, software development, and beyond. BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models. We evaluate the behavior of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters. In addition, a team of human expert raters performed all tasks in order to provide a strong baseline. Findings include: model performance and calibration both improve with scale, but are poor in absolute terms (and when compared with rater performance); performance is remarkably similar across model classes, though with benefits from sparsity; tasks that improve gradually and predictably commonly involve a large knowledge or memorization component, whereas tasks that exhibit "breakthrough" behavior at a critical scale often involve multiple steps or components, or brittle metrics; social bias typically increases with scale in settings with ambiguous context, but this can be improved with prompting.

Bryan Li, William Walden, Yu Hou et al.

Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems. Core to many evaluation frameworks is the use of atomic facts, or nuggets, to assess a report's coverage of query-relevant information attested in the underlying collection. While nuggets have traditionally been represented as short statements, recent work has used question-answer (QA) representations, enabling fine-grained evaluations that decouple the information need (i.e. the question) from the potentially diverse content that satisfies it (i.e. its answers). A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs. This challenge is acute in cross-lingual settings, where information is found in multilingual source documents. Accordingly, we introduce DoGMaTiQ, a pipeline for generating high-quality QA-based nugget sets in three stages: (1) document-grounded nugget generation, (2) paraphrase clustering, and (3) nugget subselection based on principled quality criteria. We integrate DoGMaTiQ nuggets with AutoArgue -- a recent nugget-based evaluation framework -- to enable fully automatic evaluation of generated reports. We conduct extensive experiments on two cross-lingual TREC shared tasks, NeuCLIR and RAGTIME, showing strong rank correlations with both human-in-the-loop and fully manual judgments. Finally, detailed analysis of our pipeline reveals that a strong LLM nugget generator is key, and that the system rankings induced by DoGMaTiQ are robust to outlier systems. We facilitate future research in report evaluation by publicly releasing our code and artifacts at https://github.com/manestay/dogmatiq.

Meida Chen, Qingyong Hu, Zifan Yu et al.

Although various 3D datasets with different functions and scales have been proposed recently, it remains challenging for individuals to complete the whole pipeline of large-scale data collection, sanitization, and annotation. Moreover, the created datasets usually suffer from extremely imbalanced class distribution or partial low-quality data samples. Motivated by this, we explore the procedurally synthetic 3D data generation paradigm to equip individuals with the full capability of creating large-scale annotated photogrammetry point clouds. Specifically, we introduce a synthetic aerial photogrammetry point clouds generation pipeline that takes full advantage of open geospatial data sources and off-the-shelf commercial packages. Unlike generating synthetic data in virtual games, where the simulated data usually have limited gaming environments created by artists, the proposed pipeline simulates the reconstruction process of the real environment by following the same UAV flight pattern on different synthetic terrain shapes and building densities, which ensure similar quality, noise pattern, and diversity with real data. In addition, the precise semantic and instance annotations can be generated fully automatically, avoiding the expensive and time-consuming manual annotation. Based on the proposed pipeline, we present a richly-annotated synthetic 3D aerial photogrammetry point cloud dataset, termed STPLS3D, with more than 16 $km^2$ of landscapes and up to 18 fine-grained semantic categories. For verification purposes, we also provide a parallel dataset collected from four areas in the real environment. Extensive experiments conducted on our datasets demonstrate the effectiveness and quality of the proposed synthetic dataset.

Te-Lin Wu, Zi-Yi Dou, Qingyuan Hu et al.

Multimodal counterfactual reasoning is a vital yet challenging ability for AI systems. It involves predicting the outcomes of hypothetical circumstances based on vision and language inputs, which enables AI models to learn from failures and explore hypothetical scenarios. Despite its importance, there are only a few datasets targeting the counterfactual reasoning abilities of multimodal models. Among them, they only cover reasoning over synthetic environments or specific types of events (e.g. traffic collisions), making them hard to reliably benchmark the model generalization ability in diverse real-world scenarios and reasoning dimensions. To overcome these limitations, we develop a video question answering dataset, ACQUIRED: it consists of 3.9K annotated videos, encompassing a wide range of event types and incorporating both first and third-person viewpoints, which ensures a focus on real-world diversity. In addition, each video is annotated with questions that span three distinct dimensions of reasoning, including physical, social, and temporal, which can comprehensively evaluate the model counterfactual abilities along multiple aspects. We benchmark our dataset against several state-of-the-art language-only and multimodal models and experimental results demonstrate a significant performance gap (>13%) between models and humans. The findings suggest that multimodal counterfactual reasoning remains an open challenge and ACQUIRED is a comprehensive and reliable benchmark for inspiring future research in this direction.

Jiao Sun, Yu Hou, Jiin Kim et al.

Information-seeking AI assistant systems aim to answer users' queries about knowledge in a timely manner. However, both the human-perceived helpfulness of information-seeking assistant systems and its fairness implication are under-explored. In this paper, we study computational measurements of helpfulness. We collect human annotations on the helpfulness of dialogue responses, develop models for automatic helpfulness evaluation, and then propose to use the helpfulness level of a dialogue system towards different user queries to gauge the fairness of a dialogue system. Experiments with state-of-the-art dialogue systems, including ChatGPT, under three information-seeking scenarios reveal that existing systems tend to be more helpful for questions regarding concepts from highly-developed countries than less-developed countries, uncovering potential fairness concerns underlying the current information-seeking assistant systems.

Maharshi Gor, Yoo Yeon Sung, Yu Hou et al.

AI systems are fallible, and humans can make mistakes in deciding whether to trust AI over their own judgment. Thus, improving human-AI collaboration requires understanding when, why, and how humans decide to rely on AI. We study two distinct reliance decisions: the delegation choice -- deciding when to let AI act autonomously without knowing its output, and the adoption choice -- evaluating AI suggestions and deciding how to use them. Both of these decoupled reliance patterns shape collaboration, but prior work rarely studies them together in realistic settings with the same users. We address this gap by studying collaborative human--AI teams competing in a question-answering game in which humans can choose when and how to work with AI agents to win. Our 24 matches pair 23 expert humans with 16 AI agents, capturing 387 delegation and 1440 adoption decisions. While human--AI collaboration performs better than either AI or humans alone, humans make suboptimal collaboration decisions, both under-relying on correct AI suggestions (3.9% of opportunities missed) and over-relying when AI misleads them (1.7%). Both parties contribute wrong answers: reported model confidence is near chance when humans and AI disagree, while confirmation bias drives higher under-reliance (64.5%) when an AI suggestion agrees with humans' initial incorrect answer. To close this gap, we recommend calibrated confidence, evidence-grounded explanations, and mechanisms that help users refine trust.

Zaifu Zhan, Min Zeng, Shuang Zhou et al.

Objective: To improve the efficiency of medical question answering (MedQA) with large language models (LLMs) by avoiding unnecessary reasoning while maintaining accuracy. Methods: We propose Selective Chain-of-Thought (Selective CoT), an inference-time strategy that first predicts whether a question requires reasoning and generates a rationale only when needed. Two open-source LLMs (Llama-3.1-8B and Qwen-2.5-7B) were evaluated on four biomedical QA benchmarks-HeadQA, MedQA-USMLE, MedMCQA, and PubMedQA. Metrics included accuracy, total generated tokens, and inference time. Results: Selective CoT reduced inference time by 13-45% and token usage by 8-47% with minimal accuracy loss ($\leq$4\%). In some model-task pairs, it achieved both higher accuracy and greater efficiency than standard CoT. Compared with fixed-length CoT, Selective CoT reached similar or superior accuracy at substantially lower computational cost. Discussion: Selective CoT dynamically balances reasoning depth and efficiency by invoking explicit reasoning only when beneficial, reducing redundancy on recall-type questions while preserving interpretability. Conclusion: Selective CoT provides a simple, model-agnostic, and cost-effective approach for medical QA, aligning reasoning effort with question complexity to enhance real-world deployability of LLM-based clinical systems.

Yu Hou, Cong Tran, Won-Yong Shin

The discovery of community structures in social networks has gained considerable attention as a fundamental problem for various network analysis tasks. However, due to privacy concerns or access restrictions, the network structure is often unknown, thereby rendering established community detection approaches ineffective without costly data acquisition. To tackle this challenge, we present META-CODE, a novel end-to-end solution for detecting overlapping communities in networks with unknown topology via exploratory learning aided by easy-to-collect node metadata. Specifically, META-CODE consists of three steps: 1) initial network inference, 2) node-level community-affiliation embedding based on graph neural networks (GNNs) trained by our new reconstruction loss, and 3) network exploration via community-affiliation-based node queries, where Steps 2 and 3 are performed iteratively. Experimental results demonstrate that META-CODE exhibits (a) superiority over benchmark methods for overlapping community detection, (b) the effectiveness of our training model, and (c) fast network exploration.

Ying Liu, Haozhu Wang, Huixue Zhou et al.

Reinforcement learning (RL) has emerged as a powerful approach for tackling complex medical decision-making problems such as treatment planning, personalized medicine, and optimizing the scheduling of surgeries and appointments. It has gained significant attention in the field of Natural Language Processing (NLP) due to its ability to learn optimal strategies for tasks such as dialogue systems, machine translation, and question-answering. This paper presents a review of the RL techniques in NLP, highlighting key advancements, challenges, and applications in healthcare. The review begins by visualizing a roadmap of machine learning and its applications in healthcare. And then it explores the integration of RL with NLP tasks. We examined dialogue systems where RL enables the learning of conversational strategies, RL-based machine translation models, question-answering systems, text summarization, and information extraction. Additionally, ethical considerations and biases in RL-NLP systems are addressed.

Yu Hou, Cong Tran, Ming Li et al.

In social networks, the discovery of community structures has received considerable attention as a fundamental problem in various network analysis tasks. However, due to privacy concerns or access restrictions, the network structure is often uncertain, thereby rendering established community detection approaches ineffective without costly network topology acquisition. To tackle this challenge, we present META-CODE, a unified framework for detecting overlapping communities via exploratory learning aided by easy-to-collect node metadata when networks are topologically unknown (or only partially known). Specifically, META-CODE consists of three iterative steps in addition to the initial network inference step: 1) node-level community-affiliation embeddings based on graph neural networks (GNNs) trained by our new reconstruction loss, 2) network exploration via community-affiliation-based node queries, and 3) network inference using an edge connectivity-based Siamese neural network model from the explored network. Through extensive experiments on three real-world datasets including two large networks, we demonstrate: (a) the superiority of META-CODE over benchmark community detection methods, achieving remarkable gains up to 65.55% on the Facebook dataset over the best competitor among our selected competitive methods in terms of normalized mutual information (NMI), (b) the impact of each module in META-CODE, (c) the effectiveness of node queries in META-CODE based on empirical evaluations and theoretical findings, and (d) the convergence of the inferred network.

Zaifu Zhan, Yu Hou, Kai Yu et al.

Rare diseases affect over 300 million people worldwide and are characterized by complex care pathways, limited clinical expertise, and substantial unmet communication needs throughout the long patient journey. Recent advances in large language models (LLMs) offer new opportunities to support patient education and communication, yet their application in rare diseases remains unclear. We conducted a scoping review of studies published between January 2022 and March 2026 across major databases, identifying 12 studies on LLM-based rare disease patient education and communication. Data were extracted on study characteristics, application scenarios, model usage, and evaluation methods, and synthesized using descriptive and qualitative analyses. The literature is highly recent and dominated by general-purpose models, particularly ChatGPT. Most studies focus on patient question answering using curated question sets, with limited use of real-world data or longitudinal communication scenarios. Evaluations are primarily centered on accuracy, with limited attention to patient-centered dimensions such as readability, empathy, and communication quality. Multilingual communication is rarely addressed. Overall, the field remains at an early stage. Future research should prioritize patient-centered design, domain-adapted methods, and real-world deployment to support safe, adaptive, and effective communication in rare diseases.

Shuang Zhou, Jiashuo Wang, Zidu Xu et al.

Explainable disease diagnosis, which leverages patient information (e.g., signs and symptoms) and computational models to generate probable diagnoses and reasonings, offers clear clinical values. However, when clinical notes encompass insufficient evidence for a definite diagnosis, such as the absence of definitive symptoms, diagnostic uncertainty usually arises, increasing the risk of misdiagnosis and adverse outcomes. Although explicitly identifying and explaining diagnostic uncertainties is essential for trustworthy diagnostic systems, it remains under-explored. To fill this gap, we introduce ConfiDx, an uncertainty-aware large language model (LLM) created by fine-tuning open-source LLMs with diagnostic criteria. We formalized the task and assembled richly annotated datasets that capture varying degrees of diagnostic ambiguity. Evaluating ConfiDx on real-world datasets demonstrated that it excelled in identifying diagnostic uncertainties, achieving superior diagnostic performance, and generating trustworthy explanations for diagnoses and uncertainties. To our knowledge, this is the first study to jointly address diagnostic uncertainty recognition and explanation, substantially enhancing the reliability of automatic diagnostic systems.

William Walden, Marc Mason, Orion Weller et al.

Generation of long-form, citation-backed reports is a primary use case for retrieval augmented generation (RAG) systems. While open-source evaluation tools exist for various RAG tasks, ones tailored to report generation (RG) are lacking. Accordingly, we introduce Auto-ARGUE, a robust LLM-based implementation of the recently proposed ARGUE framework for RG evaluation. We present analysis of Auto-ARGUE on the RG pilot task from the TREC 2024 NeuCLIR track, showing good system-level correlations with human judgments. We further release a web app for visualization of Auto-ARGUE outputs.

Jiaxin Yang, Yu Hou, Muxin Liu et al.

Can general-purpose image editors predict physical maps from a single RGB image? General-purpose image editors differ from standard task-specific dense-prediction models: they do not directly take an image and output a physical map. Instead, they must be guided by prompts, examples, or image-based textual cues. To this end, we introduce PhysEditBench, a novel protocol-conditioned benchmark to evaluate and standardize image editors in dense physical-map prediction that covers five targets: depth, normal, albedo, roughness, and metallic maps. For evaluation data, we build a target-dependent benchmark substrate. We use OpenRooms-FF for depth, surface normal, albedo, and roughness, InteriorVerse as an additional source for depth, normal, albedo, and a new procedurally generated source for metallic maps. We curate the data with quality checks, valid-region masks, scene-level sampling, and lighting-based stress subsets to ensure reliable and diverse evaluation. For each target, PhysEditBench defines a fixed protocol that specifies the allowed input, expected output format, and scoring procedure. Each score, therefore, reflects the performance of a model under a specified protocol, rather than its best possible performance under all prompts or interaction modes. Experimental results show that specialized models remain much stronger on depth, normal, and albedo, and stronger image editors can produce more reasonable map-like outputs. For roughness and metallic, image editors can match or outperform specialized baselines on some scalar metrics, but they still suffer from structural errors, sparsity effects, and sensitivity to lighting.

Jia Li, Yu Hou, Rui Zhang

Automatically discovering personalized sequential events from large-scale time-series data is crucial for enabling precision medicine in clinical research, yet it remains a formidable challenge even for contemporary AI models. For example, while transformers capture rich associations, they are mostly agnostic to event timing and ordering, thereby bypassing potential causal reasoning. Intuitively, we need a method capable of evaluating the "degree of alignment" among patient-specific trajectories and identifying their shared patterns, i.e., the significant events in a consistent sequence. This necessitates treating timing as a true \emph{computable} dimension, allowing models to assign ``relative timestamps'' to candidate events beyond their observed physical times. In this work, we introduce LITT, a novel Timing-Transformer architecture that enables temporary alignment of sequential events on a virtual ``relative timeline'', thereby enabling \emph{event-timing-focused attention} and personalized interpretations of clinical trajectories. Its interpretability and effectiveness are validated on real-world longitudinal EHR data from 3,276 breast cancer patients to predict the onset timing of cardiotoxicity-induced heart disease. Furthermore, LITT outperforms both the benchmark and state-of-the-art survival analysis methods on public datasets, positioning it as a significant step forward for precision medicine in clinical AI.

Shikhar Singh, Nuan Wen, Yu Hou et al.

Commonsense reasoning is intuitive for humans but has been a long-term challenge for artificial intelligence (AI). Recent advancements in pretrained language models have shown promising results on several commonsense benchmark datasets. However, the reliability and comprehensiveness of these benchmarks towards assessing model's commonsense reasoning ability remains unclear. To this end, we introduce a new commonsense reasoning benchmark dataset comprising natural language true/false statements, with each sample paired with its complementary counterpart, resulting in 4k sentence pairs. We propose a pairwise accuracy metric to reliably measure an agent's ability to perform commonsense reasoning over a given situation. The dataset is crowdsourced and enhanced with an adversarial model-in-the-loop setup to incentivize challenging samples. To facilitate a systematic analysis of commonsense capabilities, we design our dataset along the dimensions of knowledge domains, reasoning scenarios and numeracy. Experimental results demonstrate that our strongest baseline (UnifiedQA-3B), after fine-tuning, achieves ~71% standard accuracy and ~51% pairwise accuracy, well below human performance (~95% for both metrics). The dataset is available at https://github.com/PlusLabNLP/Com2Sense.

Zixu Zhao, Sijin Wang, Yu Hou et al.

Tourism is a high-stakes setting for conversational recommender systems (CRS): a plausible-sounding suggestion can waste real money and trip time once a traveler acts on it. Existing CRS benchmarks primarily evaluate systems with a single Recall@k score over entity mentions, and tourism-specific resources add spatial or knowledge-graph context, yet none of them couple multi-turn recommendation with verbatim review-span evidence and rejection recovery. This leaves an evaluation gap for tourism recommendation that is simultaneously trustworthy, verifiable, and adaptive: recommend the right point of interest (POI) for multi-aspect preferences (such as cuisine, price, atmosphere, walking distance), justify each suggestion with verifiable evidence from prior visitors so the traveler can act without trial and error, and recover when the first recommendation is rejected mid-dialogue. We introduce TRACE, where each item is a multi-turn tourism recommendation dialogue with review-span citations and explicit rejection turns: 10,000 dialogues over 2,400 Yelp POIs and 34,208 reviews across eight U.S. cities, paired with 14 retrieval, planning, and LLM baselines, along with 25 metrics organized under Accuracy, Grounding, and Recovery. Across these baselines, TRACE reveals the Three-Competency Gap: LLM Zero-Shot leads in closed-set Recall@1 and rejection recovery but cites less densely than retrievers; non-LLM retrievers achieve surface-verbatim grounding but with low accuracy; Multi-Review Synthesis fails at recovery. The Grounding Score agrees with human citation precision (Spearman rho=+0.80, p<10^-20), and paired t-tests reproduce the per-baseline ranking (p<0.01 on the dominant contrasts). TRACE reframes accountable tourism recommendation as a joint target (right POI, verifiable evidence, adaptive repair) rather than a single-axis leaderboard.

Yu Hou, Jin-Duk Park, Won-Yong Shin

A recent study has shown that diffusion models are well-suited for modeling the generative process of user-item interactions in recommender systems due to their denoising nature. However, existing diffusion model-based recommender systems do not explicitly leverage high-order connectivities that contain crucial collaborative signals for accurate recommendations. Addressing this gap, we propose CF-Diff, a new diffusion model-based collaborative filtering (CF) method, which is capable of making full use of collaborative signals along with multi-hop neighbors. Specifically, the forward-diffusion process adds random noise to user-item interactions, while the reverse-denoising process accommodates our own learning model, named cross-attention-guided multi-hop autoencoder (CAM-AE), to gradually recover the original user-item interactions. CAM-AE consists of two core modules: 1) the attention-aided AE module, responsible for precisely learning latent representations of user-item interactions while preserving the model's complexity at manageable levels, and 2) the multi-hop cross-attention module, which judiciously harnesses high-order connectivity information to capture enhanced collaborative signals. Through comprehensive experiments on three real-world datasets, we demonstrate that CF-Diff is (a) Superior: outperforming benchmark recommendation methods, achieving remarkable gains up to 7.29% compared to the best competitor, (b) Theoretically-validated: reducing computations while ensuring that the embeddings generated by our model closely approximate those from the original cross-attention, and (c) Scalable: proving the computational efficiency that scales linearly with the number of users or items.

Dayeon Ki, Yu Hou, Rachel Rudinger et al.

Neologisms and emerging slang are central to daily conversation, yet challenging for non-native speakers (NNS) to interpret and use appropriately in cross-cultural communication with native speakers (NS). NNS increasingly make use of Artificial Intelligence (AI) tools to learn these words. We study the utility of such tools in mediating an informal communication scenario through a human-subjects study (N=234): NNS participants learn English neologisms with AI support, write messages using the learned word to an NS friend, and judge contextual appropriateness of the neologism in two provided writing samples. Using both NS evaluator-rated communicative competence of NNS-produced writing and NNS' contextual appropriateness judgments, we compare three AI-based support conditions: AI Definition, AI Rewrite into simpler English, AI Explanation of meaning and usage, and Non-AI Dictionary for comparison. We show that AI Explanation yields the largest gains over no support in NS-rated competence, while contextual appropriateness judgments show indifference across support. NNS participants' self-reported perceptions tend to overestimate NS ratings, revealing a mismatch between perceived and actual competence. We further observe a significant gap between NNS- and NS-produced writing, highlighting the limitations of current AI tools and informing design for future tools.

Yu Hou, Hal Daumé, Rachel Rudinger

Studies of human psychology have demonstrated that people are more motivated to extend empathy to in-group members than out-group members (Cikara et al., 2011). In this study, we investigate how this aspect of intergroup relations in humans is replicated by LLMs in an emotion intensity prediction task. In this task, the LLM is given a short description of an experience a person had that caused them to feel a particular emotion; the LLM is then prompted to predict the intensity of the emotion the person experienced on a numerical scale. By manipulating the group identities assigned to the LLM's persona (the "perceiver") and the person in the narrative (the "experiencer"), we measure how predicted emotion intensities differ between in-group and out-group settings. We observe that LLMs assign higher emotion intensity scores to in-group members than out-group members. This pattern holds across all three types of social groupings we tested: race/ethnicity, nationality, and religion. We perform an in-depth analysis on Llama-3.1-8B, the model which exhibited strongest intergroup bias among those tested.

Yu Hou, Hua Li, Ha Young Kim et al.

Diffusion models recently emerged as a powerful paradigm for recommender systems, offering state-of-the-art performance by modeling the generative process of user-item interactions. However, training such models from scratch is both computationally expensive and yields diminishing returns once convergence is reached. To remedy these challenges, we propose ReFiT, a new framework that integrates Reinforcement learning (RL)-based Fine-Tuning into diffusion-based recommender systems. In contrast to prior RL approaches for diffusion models depending on external reward models, ReFiT adopts a task-aligned design: it formulates the denoising trajectory as a Markov decision process (MDP) and incorporates a collaborative signal-aware reward function that directly reflects recommendation quality. By tightly coupling the MDP structure with this reward signal, ReFiT empowers the RL agent to exploit high-order connectivity for fine-grained optimization, while avoiding the noisy or uninformative feedback common in naive reward designs. Leveraging policy gradient optimization, ReFiT maximizes exact log-likelihood of observed interactions, thereby enabling effective post hoc fine-tuning of diffusion recommenders. Comprehensive experiments on wide-ranging real-world datasets demonstrate that the proposed ReFiT framework (a) exhibits substantial performance gains over strong competitors (up to 36.3% on sequential recommendation), (b) demonstrates strong efficiency with linear complexity in the number of users or items, and (c) generalizes well across multiple diffusion-based recommendation scenarios. The source code and datasets are publicly available at https://anonymous.4open.science/r/ReFiT-4C60.

Zaifu Zhan, Shuang Zhou, Xiaoshan Zhou et al.

Objectives: We aim to dynamically retrieve informative demonstrations, enhancing in-context learning in multimodal large language models (MLLMs) for disease classification. Methods: We propose a Retrieval-Augmented In-Context Learning (RAICL) framework, which integrates retrieval-augmented generation (RAG) and in-context learning (ICL) to adaptively select demonstrations with similar disease patterns, enabling more effective ICL in MLLMs. Specifically, RAICL examines embeddings from diverse encoders, including ResNet, BERT, BioBERT, and ClinicalBERT, to retrieve appropriate demonstrations, and constructs conversational prompts optimized for ICL. We evaluated the framework on two real-world multi-modal datasets (TCGA and IU Chest X-ray), assessing its performance across multiple MLLMs (Qwen, Llava, Gemma), embedding strategies, similarity metrics, and varying numbers of demonstrations. Results: RAICL consistently improved classification performance. Accuracy increased from 0.7854 to 0.8368 on TCGA and from 0.7924 to 0.8658 on IU Chest X-ray. Multi-modal inputs outperformed single-modal ones, with text-only inputs being stronger than images alone. The richness of information embedded in each modality will determine which embedding model can be used to get better results. Few-shot experiments showed that increasing the number of retrieved examples further enhanced performance. Across different similarity metrics, Euclidean distance achieved the highest accuracy while cosine similarity yielded better macro-F1 scores. RAICL demonstrated consistent improvements across various MLLMs, confirming its robustness and versatility. Conclusions: RAICL provides an efficient and scalable approach to enhance in-context learning in MLLMs for multimodal disease classification.

Zaifu Zhan, Shuang Zhou, Huixue Zhou et al.

The advancement of Large Language Models (LLMs) has significantly impacted biomedical Natural Language Processing (NLP), enhancing tasks such as named entity recognition, relation extraction, event extraction, and text classification. In this context, the DeepSeek series of models have shown promising potential in general NLP tasks, yet their capabilities in the biomedical domain remain underexplored. This study evaluates multiple DeepSeek models (Distilled-DeepSeek-R1 series and Deepseek-LLMs) across four key biomedical NLP tasks using 12 datasets, benchmarking them against state-of-the-art alternatives (Llama3-8B, Qwen2.5-7B, Mistral-7B, Phi-4-14B, Gemma-2-9B). Our results reveal that while DeepSeek models perform competitively in named entity recognition and text classification, challenges persist in event and relation extraction due to precision-recall trade-offs. We provide task-specific model recommendations and highlight future research directions. This evaluation underscores the strengths and limitations of DeepSeek models in biomedical NLP, guiding their future deployment and optimization.

Paola Cascante-Bonilla, Yu Hou, Yang Trista Cao et al.

Compositional reasoning in Vision-Language Models (VLMs) remains challenging as these models often struggle to relate objects, attributes, and spatial relationships. Recent methods aim to address these limitations by relying on the semantics of the textual description, using Large Language Models (LLMs) to break them down into subsets of questions and answers. However, these methods primarily operate on the surface level, failing to incorporate deeper lexical understanding while introducing incorrect assumptions generated by the LLM. In response to these issues, we present Caption Expansion with Contradictions and Entailments (CECE), a principled approach that leverages Natural Language Inference (NLI) to generate entailments and contradictions from a given premise. CECE produces lexically diverse sentences while maintaining their core meaning. Through extensive experiments, we show that CECE enhances interpretability and reduces overreliance on biased or superficial features. By balancing CECE along the original premise, we achieve significant improvements over previous methods without requiring additional fine-tuning, producing state-of-the-art results on benchmarks that score agreement with human judgments for image-text alignment, and achieving an increase in performance on Winoground of +19.2% (group score) and +12.9% on EqBen (group score) over the best prior work (finetuned with targeted data).

Yu Hou, Zaifu Zhan, Min Zeng et al.

Biomedical literature and clinical narratives pose multifaceted challenges for natural language understanding, from precise entity extraction and document synthesis to multi-step diagnostic reasoning. This study extends a unified benchmark to evaluate GPT-5 and GPT-4o under zero-, one-, and five-shot prompting across five core biomedical NLP tasks: named entity recognition, relation extraction, multi-label document classification, summarization, and simplification, and nine expanded biomedical QA datasets covering factual knowledge, clinical reasoning, and multimodal visual understanding. Using standardized prompts, fixed decoding parameters, and consistent inference pipelines, we assessed model performance, latency, and token-normalized cost under official pricing. GPT-5 consistently outperformed GPT-4o, with the largest gains on reasoning-intensive datasets such as MedXpertQA and DiagnosisArena and stable improvements in multimodal QA. In core tasks, GPT-5 achieved better chemical NER and ChemProt scores but remained below domain-tuned baselines for disease NER and summarization. Despite producing longer outputs, GPT-5 showed comparable latency and 30 to 50 percent lower effective cost per correct prediction. Fine-grained analyses revealed improvements in diagnosis, treatment, and reasoning subtypes, whereas boundary-sensitive extraction and evidence-dense summarization remain challenging. Overall, GPT-5 approaches deployment-ready performance for biomedical QA while offering a favorable balance of accuracy, interpretability, and economic efficiency. The results support a tiered prompting strategy: direct prompting for large-scale or cost-sensitive applications, and chain-of-thought scaffolds for analytically complex or high-stakes scenarios, highlighting the continued need for hybrid solutions where precision and factual fidelity are critical.

Yoo Yeon Sung, Eve Fleisig, Yu Hou et al.

Language models are often miscalibrated, leading to confidently incorrect answers. We introduce GRACE, a benchmark for language model calibration that incorporates comparison with human calibration. GRACE consists of question-answer pairs, in which each question contains a series of clues that gradually become easier, all leading to the same answer; models must answer correctly as early as possible as the clues are revealed. This setting permits granular measurement of model calibration based on how early, accurately, and confidently a model answers. After collecting these questions, we host live human vs. model competitions to gather 1,749 data points on human and model teams' timing, accuracy, and confidence. We propose a metric, CalScore, that uses GRACE to analyze model calibration errors and identify types of model miscalibration that differ from human behavior. We find that although humans are less accurate than models, humans are generally better calibrated. Since state-of-the-art models struggle on GRACE, it effectively evaluates progress on improving model calibration.

Zihao Wang, Yu Hou, Lucio Soibelman

The potential energy loss of aging buildings traps building owners in a cycle of underfunding operations and overpaying maintenance costs. Energy auditors intending to generate an energy model of a target building for performance assessment may struggle to obtain accurate results as the spatial distribution of temperatures is not considered when calculating the U-value of the building envelope. This paper proposes a pixel-level method based on infrared thermography (IRT) that considers two-dimensional (2D) spatial temperature distributions of the outdoor and indoor surfaces of the target wall to generate a 2D U-value map of the wall. The result supports that the proposed method can better reflect the actual thermal insulation performance of the target wall compared to the current IRT-based methods that use a single-point room temperature as input.

Baixiao Huang, Baiyu Huang, Yu Hou

Quadruped robots are employed in various scenarios in building construction. However, autonomous stair climbing across different indoor staircases remains a major challenge for robot dogs to complete building construction tasks. In this project, we employed a two-stage end-to-end deep reinforcement learning (RL) approach to optimize a robot's performance on U-shaped stairs. The training robot-dog modality, Unitree Go2, was first trained to climb stairs on Isaac Lab's pyramid-stair terrain, and then to climb a U-shaped indoor staircase using the learned policies. This project explores end-to-end RL methods that enable robot dogs to autonomously climb stairs. The results showed (1) the successful goal reached for robot dogs climbing U-shaped stairs with a stall penalty, and (2) the transferability from the policy trained on U-shaped stairs to deployment on straight, L-shaped, and spiral stair terrains, and transferability from other stair models to deployment on U-shaped terrain.

Baixiao Huang, Baiyu Huang, Yu Hou

Quadruped robots are used for primary searches during the early stages of indoor fires. A typical primary search involves quickly and thoroughly looking for victims under hazardous conditions and monitoring flammable materials. However, situational awareness in complex indoor environments and rapid stair climbing across different staircases remain the main challenges for robot-assisted primary searches. In this project, we designed a two-stage end-to-end deep reinforcement learning (RL) approach to optimize both navigation and locomotion. In the first stage, the quadrupeds, Unitree Go2, were trained to climb stairs in Isaac Lab's pyramid-stair terrain. In the second stage, the quadrupeds were trained to climb various realistic indoor staircases in the Isaac Lab engine, with the learned policy transferred from the previous stage. These indoor staircases are straight, L-shaped, and spiral, to support climbing tasks in complex environments. This project explores how to balance navigation and locomotion and how end-to-end RL methods can enable quadrupeds to adapt to different stair shapes. Our main contributions are: (1) A two-stage end-to-end RL framework that transfers stair-climbing skills from abstract pyramid terrain to realistic indoor stair topologies. (2) A centerline-based navigation formulation that enables unified learning of navigation and locomotion without hierarchical planning. (3) Demonstration of policy generalization across diverse staircases using only local height-map perception. (4) An empirical analysis of success, efficiency, and failure modes under increasing stair difficulty.

Yu Hou, Won-Yong Shin

Large language model (LLM)-based recommender systems have achieved high-quality performance by bridging the discrepancy between the item space and the language space through item tokenization. However, existing item tokenization methods typically require training separate models for each item domain, limiting generalization. Moreover, the diverse distributions and semantics across item domains make it difficult to construct a unified tokenization that preserves domain-specific information. To address these challenges, we propose UniTok, a Unified item Tokenization framework that integrates our own mixture-of-experts (MoE) architecture with a series of codebooks to convert items into discrete tokens, enabling scalable tokenization while preserving semantic information across multiple item domains. Specifically, items from different domains are first projected into a unified latent space through a shared encoder. They are then routed to domain-specific experts to capture the unique semantics, while a shared expert, which is always active, encodes common knowledge transferable across domains. Additionally, to mitigate semantic imbalance across domains, we present a mutual information calibration mechanism, which guides the model towards retaining similar levels of semantic information for each domain. Comprehensive experiments on wide-ranging real-world datasets demonstrate that the proposed UniTok framework is (a) highly effective: achieving up to 51.89% improvements over strong benchmarks, (b) theoretically sound: showing the analytical validity of our architectural design and optimization; and (c) highly generalizable: demonstrating robust performance across diverse domains without requiring per-domain retraining, a capability not supported by existing baselines.

Zaifu Zhan, Shuang Zhou, Min Zeng et al.

Large language models have demonstrated remarkable capabilities in biomedical natural language processing, yet their rapid growth in size and computational requirements present a major barrier to adoption in healthcare settings where data privacy precludes cloud deployment and resources are limited. In this study, we systematically evaluated the impact of quantization on 12 state-of-the-art large language models, including both general-purpose and biomedical-specific models, across eight benchmark datasets covering four key tasks: named entity recognition, relation extraction, multi-label classification, and question answering. We show that quantization substantially reduces GPU memory requirements-by up to 75%-while preserving model performance across diverse tasks, enabling the deployment of 70B-parameter models on 40GB consumer-grade GPUs. In addition, domain-specific knowledge and responsiveness to advanced prompting methods are largely maintained. These findings provide significant practical and guiding value, highlighting quantization as a practical and effective strategy for enabling the secure, local deployment of large yet high-capacity language models in biomedical contexts, bridging the gap between technical advances in AI and real-world clinical translation.

Meida Chen, Andrew Feng, Yu Hou et al.

In recent years, photogrammetry has been widely used in many areas to create photorealistic 3D virtual data representing the physical environment. The innovation of small unmanned aerial vehicles (sUAVs) has provided additional high-resolution imaging capabilities with low cost for mapping a relatively large area of interest. These cutting-edge technologies have caught the US Army and Navy's attention for the purpose of rapid 3D battlefield reconstruction, virtual training, and simulations. Our previous works have demonstrated the importance of information extraction from the derived photogrammetric data to create semantic-rich virtual environments (Chen et al., 2019). For example, an increase of simulation realism and fidelity was achieved by segmenting and replacing photogrammetric trees with game-ready tree models. In this work, we further investigated the semantic information extraction problem and focused on the ground material segmentation and object detection tasks. The main innovation of this work was that we leveraged both the original 2D images and the derived 3D photogrammetric data to overcome the challenges faced when using each individual data source. For ground material segmentation, we utilized an existing convolutional neural network architecture (i.e., 3DMV) which was originally designed for segmenting RGB-D sensed indoor data. We improved its performance for outdoor photogrammetric data by introducing a depth pooling layer in the architecture to take into consideration the distance between the source images and the reconstructed terrain model. To test the performance of our improved 3DMV, a ground truth ground material database was created using data from the One World Terrain (OWT) data repository. Finally, a workflow for importing the segmented ground materials into a virtual simulation scene was introduced, and visual results are reported in this paper.

Sunipa Dev, Emily Sheng, Jieyu Zhao et al.

Recent studies show that Natural Language Processing (NLP) technologies propagate societal biases about demographic groups associated with attributes such as gender, race, and nationality. To create interventions and mitigate these biases and associated harms, it is vital to be able to detect and measure such biases. While existing works propose bias evaluation and mitigation methods for various tasks, there remains a need to cohesively understand the biases and the specific harms they measure, and how different measures compare with each other. To address this gap, this work presents a practical framework of harms and a series of questions that practitioners can answer to guide the development of bias measures. As a validation of our framework and documentation questions, we also present several case studies of how existing bias measures in NLP -- both intrinsic measures of bias in representations and extrinsic measures of bias of downstream applications -- can be aligned with different harms and how our proposed documentation questions facilitates more holistic understanding of what bias measures are measuring.

Lige Zhang, Xiaolin Qin, Qing Li et al.

Recently, the use of digital images in various fields is increasing rapidly. To increase the number of images stored and get faster transmission of them, it is necessary to reduce the size of these images. Single bitmap block truncation coding (SBBTC) schemes are compression techniques, which are used to generate a common bitmap to quantize the R, G and B planes in color image. As one of the traditional SBBTC schemes, weighted plane (W-plane) method is famous for its simplicity and low time consumption. However, the W-plane method also has poor performance in visual quality. This paper proposes an improved SBBTC scheme based on W-plane method using parallel computing and hill climbing algorithm. Compared with various schemes, the simulation results of the proposed scheme are better than that of the reference schemes in visual quality and time consumption.