Searcharxiv⌕ Search

arXiv subjects

Shrey Jain

Publications and source records attributed to Shrey Jain.

16 recordsLinked to original sources
arXiv2026

PhylaFlow: Hybrid Flow Matching in Billera-Holmes-Vogtmann Tree Space for Phylogenetic Inference

Phylogenetic trees are hybrid objects: branch lengths vary continuously, while topologies change discretely through edge contractions and expansions. Billera-Holmes-Vogtmann (BHV) tree space provides a canonical geometry for this structure, representing each resolved topology as a Euclidean orthant and topological changes as motion across shared lower-dimensional boundaries. We introduce PhylaFlow, a hybrid flow-matching model that learns posterior-basin transport in BHV tree space. PhylaFlow is trained on BHV geodesic paths from random starting trees to short-run posterior samples, coupling continuous branch-length motion within orthants with learned boundary events and discrete topology transitions. We evaluate the learned geometry operationally: if the flow reaches posterior-relevant regions, finite-budget Bayesian refinement initialized from, or guided by, its terminal trees should recover posterior-supported topologies more efficiently. Across DS1-DS8 phylogenetic posterior benchmarks, PhylaFlow substantially reduces initial Tree-KL relative to classical initializers. After finite-budget MrBayes refinement, direct PhylaFlow improves early and intermediate topology-recovery trajectories on most datasets, while split-guided PhylaFlow-MCMC obtains the strongest hard-case results. The best PhylaFlow variant outperforms short-warmup on seven of eight datasets and PhyloGFN on five of eight under the same refinement budget. In a joint sequence-conditioned experiment, sequence embeddings steer posterior split recovery, although exact posterior topology recovery remains preliminary. These results show that hybrid flow matching can learn actionable transport in BHV tree space and provide a geometry-aware proposal mechanism for Bayesian phylogenetic inference.

q-bio.PE↗
arXiv2025

CancerGUIDE: Cancer Guideline Understanding via Internal Disagreement Estimation

The National Comprehensive Cancer Network (NCCN) provides evidence-based guidelines for cancer treatment. Translating complex patient presentations into guideline-compliant treatment recommendations is time-intensive, requires specialized expertise, and is prone to error. Advances in large language model (LLM) capabilities promise to reduce the time required to generate treatment recommendations and improve accuracy. We present an LLM agent-based approach to automatically generate guideline-concordant treatment trajectories for patients with non-small cell lung cancer (NSCLC). Our contributions are threefold. First, we construct a novel longitudinal dataset of 121 cases of NSCLC patients that includes clinical encounters, diagnostic results, and medical histories, each expertly annotated with the corresponding NCCN guideline trajectories by board-certified oncologists. Second, we demonstrate that existing LLMs possess domain-specific knowledge that enables high-quality proxy benchmark generation for both model development and evaluation, achieving strong correlation (Spearman coefficient r=0.88, RMSE = 0.08) with expert-annotated benchmarks. Third, we develop a hybrid approach combining expensive human annotations with model consistency information to create both the agent framework that predicts the relevant guidelines for a patient, as well as a meta-classifier that verifies prediction accuracy with calibrated confidence scores for treatment recommendations (AUROC=0.800), a critical capability for communicating the accuracy of outputs, custom-tailoring tradeoffs in performance, and supporting regulatory compliance. This work establishes a framework for clinically viable LLM-based guideline adherence systems that balance accuracy, interpretability, and regulatory requirements while reducing annotation costs, providing a scalable pathway toward automated clinical decision support.

cs.LG↗
arXiv2025

Demo: Healthcare Agent Orchestrator (HAO) for Patient Summarization in Molecular Tumor Boards

Molecular Tumor Boards (MTBs) are multidisciplinary forums where oncology specialists collaboratively assess complex patient cases to determine optimal treatment strategies. A central element of this process is the patient summary, typically compiled by a medical oncologist, radiation oncologist, or surgeon, or their trained medical assistant, who distills heterogeneous medical records into a concise narrative to facilitate discussion. This manual approach is often labor-intensive, subjective, and prone to omissions of critical information. To address these limitations, we introduce the Healthcare Agent Orchestrator (HAO), a Large Language Model (LLM)-driven AI agent that coordinates a multi-agent clinical workflow to generate accurate and comprehensive patient summaries for MTBs. Evaluating predicted patient summaries against ground truth presents additional challenges due to stylistic variation, ordering, synonym usage, and phrasing differences, which complicate the measurement of both succinctness and completeness. To overcome these evaluation hurdles, we propose TBFact, a ``model-as-a-judge'' framework designed to assess the comprehensiveness and succinctness of generated summaries. Using a benchmark dataset derived from de-identified tumor board discussions, we applied TBFact to evaluate our Patient History agent. Results show that the agent captured 94% of high-importance information (including partial entailments) and achieved a TBFact recall of 0.84 under strict entailment criteria. We further demonstrate that TBFact enables a data-free evaluation framework that institutions can deploy locally without sharing sensitive clinical data. Together, HAO and TBFact establish a robust foundation for delivering reliable and scalable support to MTBs.

cs.LG↗
arXiv2025

MedHELM: Holistic Evaluation of Large Language Models for Medical Tasks

While large language models (LLMs) achieve near-perfect scores on medical licensing exams, these evaluations inadequately reflect the complexity and diversity of real-world clinical practice. We introduce MedHELM, an extensible evaluation framework for assessing LLM performance for medical tasks with three key contributions. First, a clinician-validated taxonomy spanning 5 categories, 22 subcategories, and 121 tasks developed with 29 clinicians. Second, a comprehensive benchmark suite comprising 35 benchmarks (17 existing, 18 newly formulated) providing complete coverage of all categories and subcategories in the taxonomy. Third, a systematic comparison of LLMs with improved evaluation methods (using an LLM-jury) and a cost-performance analysis. Evaluation of 9 frontier LLMs, using the 35 benchmarks, revealed significant performance variation. Advanced reasoning models (DeepSeek R1: 66% win-rate; o3-mini: 64% win-rate) demonstrated superior performance, though Claude 3.5 Sonnet achieved comparable results at 40% lower estimated computational cost. On a normalized accuracy scale (0-1), most models performed strongly in Clinical Note Generation (0.73-0.85) and Patient Communication & Education (0.78-0.83), moderately in Medical Research Assistance (0.65-0.75), and generally lower in Clinical Decision Support (0.56-0.72) and Administration & Workflow (0.53-0.63). Our LLM-jury evaluation method achieved good agreement with clinician ratings (ICC = 0.47), surpassing both average clinician-clinician agreement (ICC = 0.43) and automated baselines including ROUGE-L (0.36) and BERTScore-F1 (0.44). Claude 3.5 Sonnet achieved comparable performance to top models at lower estimated cost. These findings highlight the importance of real-world, task-specific evaluation for medical use of LLMs and provides an open source framework to enable this.

cs.CL↗
arXiv2025

Universal Abstraction: Harnessing Frontier Models to Structure Real-World Data at Scale

A significant fraction of real-world patient information resides in unstructured clinical text. Medical abstraction extracts and normalizes key structured attributes from free-text clinical notes, which is the prerequisite for a variety of important downstream applications, including registry curation, clinical trial operations, and real-world evidence generation. Prior medical abstraction methods typically resort to building attribute-specific models, each of which requires extensive manual effort such as rule creation or supervised label annotation for the individual attribute, thus limiting scalability. In this paper, we show that existing frontier models already possess the universal abstraction capability for scaling medical abstraction to a wide range of clinical attributes. We present UniMedAbstractor (UMA), a unifying framework for zero-shot medical abstraction with a modular, customizable prompt template and the selection of any frontier large language models. Given a new attribute for abstraction, users only need to conduct lightweight prompt adaptation in UMA to adjust the specification in natural languages. Compared to traditional methods, UMA eliminates the need for attribute-specific training labels or handcrafted rules, thus substantially reducing the development time and cost. We conducted a comprehensive evaluation of UMA in oncology using a wide range of marquee attributes representing the cancer patient journey. These include relatively simple attributes typically specified within a single clinical note (e.g. performance status), as well as complex attributes requiring sophisticated reasoning across multiple notes at various time points (e.g. tumor staging). Based on a single frontier model such as GPT-4o, UMA matched or even exceeded the performance of state-of-the-art attribute-specific methods, each of which was tailored to the individual attribute.

cs.CL↗
arXiv2024

MedImageInsight: An Open-Source Embedding Model for General Domain Medical Imaging

In this work, we present MedImageInsight, an open-source medical imaging embedding model. MedImageInsight is trained on medical images with associated text and labels across a diverse collection of domains, including X-Ray, CT, MRI, dermoscopy, OCT, fundus photography, ultrasound, histopathology, and mammography. Rigorous evaluations demonstrate MedImageInsight's ability to achieve state-of-the-art (SOTA) or human expert level performance across classification, image-image search, and fine-tuning tasks. Specifically, on public datasets, MedImageInsight achieves SOTA in CT 3D medical image retrieval, as well as SOTA in disease classification and search for chest X-ray, dermatology, and OCT imaging. Furthermore, MedImageInsight achieves human expert performance in bone age estimation (on both public and partner data), as well as AUC above 0.9 in most other domains. When paired with a text decoder, MedImageInsight achieves near SOTA level single image report findings generation with less than 10\% the parameters of other models. Compared to fine-tuning GPT-4o with only MIMIC-CXR data for the same task, MedImageInsight outperforms in clinical metrics, but underperforms on lexical metrics where GPT-4o sets a new SOTA. Importantly for regulatory purposes, MedImageInsight can generate ROC curves, adjust sensitivity and specificity based on clinical need, and provide evidence-based decision support through image-image search (which can also enable retrieval augmented generation). In an independent clinical evaluation of image-image search in chest X-ray, MedImageInsight outperformed every other publicly available foundation model evaluated by large margins (over 6 points AUC), and significantly outperformed other models in terms of AI fairness (across age and gender). We hope releasing MedImageInsight will help enhance collective progress in medical imaging AI research and development.

eess.IV↗
arXiv2024

Personhood credentials: Artificial intelligence and the value of privacy-preserving tools to distinguish who is real online

Anonymity is an important principle online. However, malicious actors have long used misleading identities to conduct fraud, spread disinformation, and carry out other deceptive schemes. With the advent of increasingly capable AI, bad actors can amplify the potential scale and effectiveness of their operations, intensifying the challenge of balancing anonymity and trustworthiness online. In this paper, we analyze the value of a new tool to address this challenge: "personhood credentials" (PHCs), digital credentials that empower users to demonstrate that they are real people -- not AIs -- to online services, without disclosing any personal information. Such credentials can be issued by a range of trusted institutions -- governments or otherwise. A PHC system, according to our definition, could be local or global, and does not need to be biometrics-based. Two trends in AI contribute to the urgency of the challenge: AI's increasing indistinguishability from people online (i.e., lifelike content and avatars, agentic activity), and AI's increasing scalability (i.e., cost-effectiveness, accessibility). Drawing on a long history of research into anonymous credentials and "proof-of-personhood" systems, personhood credentials give people a way to signal their trustworthiness on online platforms, and offer service providers new tools for reducing misuse by bad actors. In contrast, existing countermeasures to automated deception -- such as CAPTCHAs -- are inadequate against sophisticated AI, while stringent identity verification solutions are insufficiently private for many use-cases. After surveying the benefits of personhood credentials, we also examine deployment risks and design challenges. We conclude with actionable next steps for policymakers, technologists, and standards bodies to consider in consultation with the public.

cs.CY↗
arXiv2024

Push and Pull: A Framework for Measuring Attentional Agency on Digital Platforms

We propose a framework for measuring attentional agency, which we define as a user's ability to allocate attention according to their own desires, goals, and intentions on digital platforms that use statistical learning to prioritize informational content. Such platforms extend people's limited powers of attention by extrapolating their preferences to large collections of previously unconsidered informational objects. However, platforms typically also allow users to influence the attention of other users in various ways. We introduce a formal framework for measuring how much a given platform empowers each user to both pull information into their own attention and push information into the attention of others. We also use these definitions to clarify the implications of generative foundation models and other recent advances in AI for the structure and efficiency of digital platforms. We conclude with a set of possible strategies for better understanding and reshaping attentional agency online.

cs.CY↗
arXiv2024

Verifiable evaluations of machine learning models using zkSNARKs

In a world of increasing closed-source commercial machine learning models, model evaluations from developers must be taken at face value. These benchmark results-whether over task accuracy, bias evaluations, or safety checks-are traditionally impossible to verify by a model end-user without the costly or impossible process of re-performing the benchmark on black-box model outputs. This work presents a method of verifiable model evaluation using model inference through zkSNARKs. The resulting zero-knowledge computational proofs of model outputs over datasets can be packaged into verifiable evaluation attestations showing that models with fixed private weights achieve stated performance or fairness metrics over public inputs. We present a flexible proving system that enables verifiable attestations to be performed on any standard neural network model with varying compute requirements. For the first time, we demonstrate this across a sample of real-world models and highlight key challenges and design solutions. This presents a new transparency paradigm in the verifiable evaluation of private models.

cs.LG↗
arXiv2023

Sandi: A System for Accountability and Applications in Direct Communication (Extended Abstract)

Reputation systems guide our decision making both in life and work: which restaurant to eat at, which vendor to buy from, which software dependencies to use, and who or what to trust. These systems are often based on old ideas and are failing in the face of modern threats. Fraudsters have found ways to manipulate them, undermining their integrity and utility. Generative AI adds to the problem by enabling the creation of real-looking fake narratives at scale, creating a false sense of consensus. Meanwhile, the need for reliable reputation concepts is more important than ever, as wrong decisions lead to increasingly severe outcomes: wasted time, poor service, and a feeling of injustice at best, fraud, identity theft, and ransomware at worst. In this extended abstract we introduce Sandi, a new kind of reputation system with a single well-defined purpose: to create trust through accountability in one-to-one transactions. Examples of such transactions include sending an email or making a purchase online. Sandi has strong security and privacy properties that make it suitable for use also in sensitive contexts. Furthermore, Sandi can guarantee reputation integrity and transparency for its registered users. As a primary application, we envision how Sandi could counter fraud and abuse in direct communication. Concretely, message senders request a cryptographic tag from Sandi that they send along with their message. If the receiver finds the message inappropriate, they can report the sender using this tag. Notably, only senders need registered accounts and do not need to manage long-term keys. The design of Sandi ensures compatibility with any communication system that allows for small binary data transmission.

cs.CR↗
arXiv2023

Contextual Confidence and Generative AI

Generative AI models perturb the foundations of effective human communication. They present new challenges to contextual confidence, disrupting participants' ability to identify the authentic context of communication and their ability to protect communication from reuse and recombination outside its intended context. In this paper, we describe strategies--tools, technologies and policies--that aim to stabilize communication in the face of these challenges. The strategies we discuss fall into two broad categories. Containment strategies aim to reassert context in environments where it is currently threatened--a reaction to the context-free expectations and norms established by the internet. Mobilization strategies, by contrast, view the rise of generative AI as an opportunity to proactively set new and higher expectations around privacy and authenticity in mediated communication.

cs.AI↗
arXiv2023

Unsupervised Learning of Molecular Embeddings for Enhanced Clustering and Emergent Properties for Chemical Compounds

The detailed analysis of molecular structures and properties holds great potential for drug development discovery through machine learning. Developing an emergent property in the model to understand molecules would broaden the horizons for development with a new computational tool. We introduce various methods to detect and cluster chemical compounds based on their SMILES data. Our first method, analyzing the graphical structures of chemical compounds using embedding data, employs vector search to meet our threshold value. The results yielded pronounced, concentrated clusters, and the method produced favorable results in querying and understanding the compounds. We also used natural language description embeddings stored in a vector database with GPT3.5, which outperforms the base model. Thus, we introduce a similarity search and clustering algorithm to aid in searching for and interacting with molecules, enhancing efficiency in chemical exploration and enabling future development of emergent properties in molecular property prediction models.

physics.chem-ph↗
arXiv2023

AI and Democracy's Digital Identity Crisis

AI-enabled tools have become sophisticated enough to allow a small number of individuals to run disinformation campaigns of an unprecedented scale. Privacy-preserving identity attestations can drastically reduce instances of impersonation and make disinformation easy to identify and potentially hinder. By understanding how identity attestations are positioned across the spectrum of decentralization, we can gain a better understanding of the costs and benefits of various attestations. In this paper, we discuss attestation types, including governmental, biometric, federated, and web of trust-based, and include examples such as e-Estonia, China's social credit system, Worldcoin, OAuth, X (formerly Twitter), Gitcoin Passport, and EAS. We believe that the most resilient systems create an identity that evolves and is connected to a network of similarly evolving identities that verify one another. In this type of system, each entity contributes its respective credibility to the attestation process, creating a larger, more comprehensive set of attestations. We believe these systems could be the best approach to authenticating identity and protecting against some of the threats to democracy that AI can pose in the hands of malicious actors. However, governments will likely attempt to mitigate these risks by implementing centralized identity authentication systems; these centralized systems could themselves pose risks to the democratic processes they are built to defend. We therefore recommend that policymakers support the development of standards-setting organizations for identity, provide legal clarity for builders of decentralized tooling, and fund research critical to effective identity authentication systems.

cs.CR↗
arXiv2023

Face Recognition in the age of CLIP & Billion image datasets

CLIP (Contrastive Language-Image Pre-training) models developed by OpenAI have achieved outstanding results on various image recognition and retrieval tasks, displaying strong zero-shot performance. This means that they are able to perform effectively on tasks for which they have not been explicitly trained. Inspired by the success of OpenAI CLIP, a new publicly available dataset called LAION-5B was collected which resulted in the development of open ViT-H/14, ViT-G/14 models that outperform the OpenAI L/14 model. The LAION-5B dataset also released an approximate nearest neighbor index, with a web interface for search & subset creation. In this paper, we evaluate the performance of various CLIP models as zero-shot face recognizers. Our findings show that CLIP models perform well on face recognition tasks, but increasing the size of the CLIP model does not necessarily lead to improved accuracy. Additionally, we investigate the robustness of CLIP models against data poisoning attacks by testing their performance on poisoned data. Through this analysis, we aim to understand the potential consequences and misuse of search engines built using CLIP models, which could potentially function as unintentional face recognition engines.

cs.CV↗
arXiv2022

A Plural Decentralized Identity Frontier: Abstraction v. Composability Tradeoffs in Web3

In this article, we explore the tension between abstraction and composability in web3 today, specifically within identity solutions, and argue that the current standard DID v1.0 is sufficiently under specified, allowing for many methods and instantiations, including blockchain based certificates. We view experiments today in web3 identity as additive and complementary, and argue that often cited differences are of degree and more in form, less in substance. By way of illustration, we compare decentralized naming services and blockchain based identity certificates such as soulbound tokens (SBTs) to decentralized identifiers (DIDs) and verifiable credentials (VCs). Both paradigms, to the extent they can be meaningfully differentiated, share similar potential as well as challenges. Specifically, we refer to fears about non consensual verification (scarlet letters) and show DID method iterations are not immune by issuing an innocuous public scarlet letter to a DIDs associated public address for anyone to see. Moreover, we argue that because SBTs are unspecified, one could characterize SBTs as an iteration, or extension, of VCs that additionally aspire to achieve composability with web3 smart contracts for correct execution of code, privacy, coercion resistance, and censorship resistance. We offer research paths for how VCs can also achieve these properties. We do not comment on cost, scalability, transferability, or common knowledge as they have been previously reviewed.

cs.CR↗