arXiv subjects
Publications and source records attributed to Hamid Hassanzadeh.
It is shown that in a Cohen-Macaulay local ring, the generic linkage of an ideal $I$ is a deformation of the arbitrary linkage of $I$. This fact does not need $I$ to be a Cohen-Macaulay ideal. The same holds for $s$-residual intersections of $I$ when $s$ does not exceed the height of $I$ by one. Under some slight conditions on $I$, one further generalizes this principle to encompass any $s$-residual intersection.
In-context learning (ICL) enables large language models (LLMs) to perform new tasks using only a few demonstrations. However, in Named Entity Recognition (NER), existing ICL methods typically rely on task-agnostic semantic similarity for demonstration retrieval, which often yields less relevant examples and leads to inferior results. We introduce DEER, a training-free ICL approach that enables LLMs to make more informed entity predictions through the use of label-grounded statistics. DEER leverages token-level statistics from training labels to identify tokens most informative for entity recognition, enabling entity-focused demonstrations. It further uses these statistics to detect and refine error-prone tokens through a targeted reflection step. Evaluated on five NER datasets across four LLMs, DEER consistently outperforms existing ICL methods and achieves performance comparable to supervised fine-tuning. Further analyses demonstrate that DEER improves example retrieval, remains effective on both seen and unseen entities, and exhibits strong robustness in low-resource settings.
One considers quasihomogeneous isolated singularities of hypersurfaces in arbitrary dimensions through the lenses of three apparently quite apart themes: syzygies, singularity invariants, and foliations. In the first of these, one adds to the well-known result of Saito's a syzygy-theoretic characterization of a quasihomogeneous singularity affording an effective computational criterion. In the second theme, one explores the Milnor-Tjurina difference number from a commutative algebra viewpoint. Building on the Briancon-Skoda theorem and exponent, we extend previously known inequalities by Dimca and Greuel to arbitrary dimension and provide algebraic formulas involving the syzygy-theoretic part and reduction exponents. In the last theme one recovers and bring up to an algebraic light a result of Camacho and Movasati by establishing a couple of characterizations of quasihomogeneous isolated singularities in terms of the generators of the module of invariant vector fields.
Clinical Question Answering (QA) systems enable doctors to quickly access patient information from electronic health records (EHRs). However, training these systems requires significant annotated data, which is limited due to the expertise needed and the privacy concerns associated with clinical data. This paper explores generating Clinical QA data using large language models (LLMs) in a zero-shot setting. We find that naive prompting often results in easy questions that do not reflect the complexity of clinical scenarios. To address this, we propose two prompting strategies: 1) instructing the model to generate questions that do not overlap with the input context, and 2) summarizing the input record using a predefined schema to scaffold question generation. Experiments on two Clinical QA datasets demonstrate that our method generates more challenging questions, significantly improving fine-tuning performance over baselines. We compare synthetic and gold data and find a gap between their training efficacy resulting from the quality of synthetically generated answers.
Extractive question answering (QA) systems can enable physicians and researchers to query medical records, a foundational capability for designing clinical studies and understanding patient medical history. However, building these systems typically requires expert-annotated QA pairs. Large language models (LLMs), which can perform extractive QA, depend on high quality data in their prompts, specialized for the application domain. We introduce a novel approach, XAIQA, for generating synthetic QA pairs at scale from data naturally available in electronic health records. Our method uses the idea of a classification model explainer to generate questions and answers about medical concepts corresponding to medical codes. In an expert evaluation with two physicians, our method identifies $2.2\times$ more semantic matches and $3.8\times$ more clinical abbreviations than two popular approaches that use sentence transformers to create QA pairs. In an ML evaluation, adding our QA pairs improves performance of GPT-4 as an extractive QA model, including on difficult questions. In both the expert and ML evaluations, we examine trade-offs between our method and sentence transformers for QA pair generation depending on question difficulty.