Personal large language model (LLM) agents increasingly perform tasks that require access to user data, raising concerns about appropriate data disclosure. We show that relying solely on LLMs to make data-sharing decisions is insufficient. Prompting LLMs to ground their decisions on contextual privacy norms fails to capture individual users’ privacy preferences, while providing prior user data-sharing decisions through in-context learning (ICL) leads to unreliable and opaque reasoning. To address these limitations, we propose ARIEL (Agentic Reasoning with Individualized Entailment Logic), a framework that combines LLMs with rule-based logic to enable structured, personalized privacy reasoning. The core mechanism of ARIEL determines whether a user’s prior decision on a data-sharing request logically entails the same decision for a new request. Experimental evaluations using advanced models and public datasets show that ARIEL reduces the F1 error rate for appropriate judgments by 40.6% compared to standard ICL-based reasoning, indicating that ARIEL is effective at correctly judging requests where the user would approve data sharing. These results demonstrate that integrating LLMs with logical entailment provides an effective and interpretable approach for automating personalized privacy decisions.

We study the ability of language models to reason about appropriate information disclosure - a central aspect of the evolving field of agentic privacy. Whereas previous works have focused on evaluating a model's ability to align with human decisions, we examine the role of ambiguity and missing context on model performance when making information-sharing decisions. We identify context ambiguity as a crucial barrier for high performance in privacy assessments. By designing Camber, a framework for context disambiguation, we show that model-generated decision rationales can reveal ambiguities and that systematically disambiguating context based on these rationales leads to significant accuracy improvements (up to 13.3% in precision and up to 22.3% in recall) as well as reductions in prompt sensitivity. Overall, our results indicate that approaches for context disambiguation are a promising way forward to enhance agentic privacy reasoning.

The growing use of large language model (LLM)-based conversational agents to manage sensitive user data raises significant privacy concerns.While these agents excel at understanding and acting on context, this capability can be exploited by malicious actors.We introduce a novel threat model where adversarial third-party apps manipulate the context of interaction to trick LLM-based agents into revealing private information not relevant to the task at hand.Grounded in the framework of contextual integrity, we introduce AirGapAgent, a privacy-conscious agent designed to prevent unintended data leakage by restricting the agent's access to only the data necessary for a specific task.Extensive experiments using Gemini, GPT, and Mistral models as agents validate our approach's effectiveness in mitigating this form of context hijacking while maintaining core agent functionality.For example, we show that a single-query context hijacking attack on a Gemini Ultra agent reduces its ability to protect user data from 94% to 45%, while an AirGapAgent achieves 97% protection, rendering the same attack ineffective. CCS Concepts• Security and privacy → Information flow control.

MOTIVATION: Gene regulatory networks define regulatory relationships between transcription factors and target genes within a biological system, and reconstructing them is essential for understanding cellular growth and function. Methods for inferring and reconstructing networks from genomics data have evolved rapidly over the last decade in response to advances in sequencing technology and machine learning. The scale of data collection has increased dramatically; the largest genome-wide gene expression datasets have grown from thousands of measurements to millions of single cells, and new technologies are on the horizon to increase to tens of millions of cells and above. RESULTS: In this work, we present the Inferelator 3.0, which has been significantly updated to integrate data from distinct cell types to learn context-specific regulatory networks and aggregate them into a shared regulatory network, while retaining the functionality of the previous versions. The Inferelator is able to integrate the largest single-cell datasets and learn cell-type-specific gene regulatory networks. Compared to other network inference methods, the Inferelator learns new and informative Saccharomyces cerevisiae networks from single-cell gene expression data, measured by recovery of a known gold standard. We demonstrate its scaling capabilities by learning networks for multiple distinct neuronal and glial cell types in the developing Mus musculus brain at E18 from a large (1.3 million) single-cell gene expression dataset with paired single-cell chromatin accessibility data. AVAILABILITY AND IMPLEMENTATION: The inferelator software is available on GitHub (https://github.com/flatironinstitute/inferelator) under the MIT license and has been released as python packages with associated documentation (https://inferelator.readthedocs.io/). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Machine learning models for predicting cell-type-specific transcription factor (TF) binding sites have become increasingly more accurate thanks to the increased availability of next-generation sequencing data and more standardized model evaluation criteria. However, knowledge transfer from data-rich to data-limited TFs and cell types remains crucial for improving TF binding prediction models because available binding labels are highly skewed towards a small collection of TFs and cell types. Transfer prediction of TF binding sites can potentially benefit from a multitask learning approach; however, existing methods typically use shallow single-task models to generate low-resolution predictions. Here, we propose NetTIME, a multitask learning framework for predicting cell-type-specific TF binding sites with base-pair resolution. RESULTS: We show that the multitask learning strategy for TF binding prediction is more efficient than the single-task approach due to the increased data availability. NetTIME trains high-dimensional embedding vectors to distinguish TF and cell-type identities. We show that this approach is critical for the success of the multitask learning strategy and allows our model to make accurate transfer predictions within and beyond the training panels of TFs and cell types. We additionally train a linear-chain conditional random field (CRF) to classify binding predictions and show that this CRF eliminates the need for setting a probability threshold and reduces classification noise. We compare our method's predictive performance with two state-of-the-art methods, Catchitt and Leopard, and show that our method outperforms previous methods under both supervised and transfer learning settings. AVAILABILITY AND IMPLEMENTATION: NetTIME is freely available at https://github.com/ryi06/NetTIME and the code is also archived at https://doi.org/10.5281/zenodo.6994897. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.