arXiv subjects

Andrei Irimia

Publications and source records attributed to Andrei Irimia.

7 recordsLinked to original sources

arXiv2026

Graph Neural Network Reveals the Cortical Morphology of Local Brain Aging in Normal Cognition and Alzheimer's Disease

Estimating brain age (BA) from T1-weighted magnetic resonance images (MRIs) provides a powerful framework for quantifying anatomical brain aging. Whereas global BA (GBA) summarizes overall brain health, local BA (LBA) provides cortically specific patterns of aging at the subject level. Although previous studies have examined anatomical contributors to GBA, to our knowledge, no framework has been established to estimate LBA using cortical morphology. To address this gap, we introduce a graph neural network (GNN) that uses morphometric features$\unicode{x2013}$cortical thickness, surface area, curvature, gray/white matter intensity ratio (GWR), sulcal depth$\unicode{x2013}$to estimate LBA across the cortical surface at high spatial resolution (mean inter-vertex distance = 1.37 mm). Trained on cortical surface meshes extracted from the MRIs of cognitively normal (CN) adults (N = 14,423), our model achieves lower mean absolute error (MAE) than the existing state-of-the-art while identifying more biologically plausible patterns of aging in Alzheimer's disease (AD) on the ADNI dataset. Association cortices emerge as primary sites of morphometric aging in CNs, whereas mild cognitive impairment is characterized by widespread aging that is pronounced in the parahippocampal gyrus. AD subjects demonstrate significant aging across the entire cortex, particularly within medial temporal regions and associated cortical networks. Feature ablation highlights curvature and GWR as preferentially sensitive to AD pathology. Regional LBA gaps are significantly associated with neuropsychological measures of AD-related cognitive impairment, linking cortical aging patterns to clinical outcomes. These results demonstrate that GNN-based modeling of cortical morphometry enables biologically interpretable mapping of local brain aging with greater interpretability than prior work.

q-bio.NC↗

arXiv2025

HDLCoRe: A Training-Free Framework for Mitigating Hallucinations in LLM-Generated HDL

Recent advances in large language models (LLMs) have demonstrated remarkable capabilities in code generation tasks. However, when applied to hardware description languages (HDL), these models exhibit significant limitations due to data scarcity, resulting in hallucinations and incorrect code generation. To address these challenges, we propose HDLCoRe, a training-free framework that enhances LLMs' HDL generation capabilities through prompt engineering techniques and retrieval-augmented generation (RAG). Our approach consists of two main components: (1) an HDL-aware Chain-of-Thought (CoT) prompting technique with self-verification that classifies tasks by complexity and type, incorporates domain-specific knowledge, and guides LLMs through step-by-step self-simulation for error correction; and (2) a two-stage heterogeneous RAG system that addresses formatting inconsistencies through key component extraction and efficiently retrieves relevant HDL examples through sequential filtering and re-ranking. HDLCoRe eliminates the need for model fine-tuning while substantially improving LLMs' HDL generation capabilities. Experimental results demonstrate that our framework achieves superior performance on the RTLLM2.0 benchmark, significantly reducing hallucinations and improving both syntactic and functional correctness.

cs.CL↗

arXiv2025

Exploiting Application-to-Architecture Dependencies for Designing Scalable OS

With the advent of hundreds of cores on a chip to accelerate applications, the operating system (OS) needs to exploit the existing parallelism provided by the underlying hardware resources to determine the right amount of processes to be mapped on the multi-core systems. However, the existing OS is not scalable and is oblivious to applications. We address these issues by adopting a multi-layer network representation of the dynamic application-to OS-to-architecture dependencies, namely the NetworkedOS. We adopt a compile-time analysis and construct a network representing the dependencies between dynamic instructions translated from the applications and the kernel and services. We propose an overlapping partitioning scheme to detect the clusters or processes that can potentially run in parallel to be mapped onto cores while reducing the number of messages transferred. At run time, processes are mapped onto the multi-core systems, taking into consideration the process affinity. Our experimental results indicate that NetworkedOS achieves performance improvement as high as 7.11x compared to Linux running on a 128-core system and 2.01x to Barrelfish running on a 64-core system.

cs.OS↗

arXiv2024

ICML Topological Deep Learning Challenge 2024: Beyond the Graph Domain

This paper describes the 2nd edition of the ICML Topological Deep Learning Challenge that was hosted within the ICML 2024 ELLIS Workshop on Geometry-grounded Representation Learning and Generative Modeling (GRaM). The challenge focused on the problem of representing data in different discrete topological domains in order to bridge the gap between Topological Deep Learning (TDL) and other types of structured datasets (e.g. point clouds, graphs). Specifically, participants were asked to design and implement topological liftings, i.e. mappings between different data structures and topological domains --like hypergraphs, or simplicial/cell/combinatorial complexes. The challenge received 52 submissions satisfying all the requirements. This paper introduces the main scope of the challenge, and summarizes the main results and findings.

cs.LG↗

arXiv2023

Leader-Follower Neural Networks with Local Error Signals Inspired by Complex Collectives

The collective behavior of a network with heterogeneous, resource-limited information processing units (e.g., group of fish, flock of birds, or network of neurons) demonstrates high self-organization and complexity. These emergent properties arise from simple interaction rules where certain individuals can exhibit leadership-like behavior and influence the collective activity of the group. Motivated by the intricacy of these collectives, we propose a neural network (NN) architecture inspired by the rules observed in nature's collective ensembles. This NN structure contains workers that encompass one or more information processing units (e.g., neurons, filters, layers, or blocks of layers). Workers are either leaders or followers, and we train a leader-follower neural network (LFNN) by leveraging local error signals and optionally incorporating backpropagation (BP) and global loss. We investigate worker behavior and evaluate LFNNs through extensive experimentation. Our LFNNs trained with local error signals achieve significantly lower error rates than previous BP-free algorithms on MNIST and CIFAR-10 and even surpass BP-enabled baselines. In the case of ImageNet, our LFNN-l demonstrates superior scalability and outperforms previous BP-free algorithms by a significant margin.

cs.LG↗

arXiv2023

Linking Symptom Inventories using Semantic Textual Similarity

An extensive library of symptom inventories has been developed over time to measure clinical symptoms, but this variety has led to several long standing issues. Most notably, results drawn from different settings and studies are not comparable, which limits reproducibility. Here, we present an artificial intelligence (AI) approach using semantic textual similarity (STS) to link symptoms and scores across previously incongruous symptom inventories. We tested the ability of four pre-trained STS models to screen thousands of symptom description pairs for related content - a challenging task typically requiring expert panels. Models were tasked to predict symptom severity across four different inventories for 6,607 participants drawn from 16 international data sources. The STS approach achieved 74.8% accuracy across five tasks, outperforming other models tested. This work suggests that incorporating contextual, semantic information can assist expert decision-making processes, yielding gains for both general and disease-specific clinical assessment.

cs.CL↗

arXiv2005

Acceptances for Space-Based and Ground-Based Fluorescence Detectors, and Inference of the Neutrino-Nucleon Cross-Section above 10^19 eV

Detection of ultra-high energy neutrinos will be useful for unraveling the dynamics of the most violent sources in the cosmos and for revealing the neutrino cross-section at extreme energy. Above ~ 10^20 eV, neutrinos may well be the only cosmic primariies. Thus, it is important to know the acceptances (event rate/flux) of proposed air-shower experiments for detecting ``horizontal'' neutrino events initiated in our atmosphere, and ``Earth-skimming'' events initiated in the Earth's surface rock or ocean. We calculate these acceptances for fluorescence detectors, both space-based as with the EUSO and OWL proposals, and ground-based, as with Auger, HiRes and Telescope Array. The neutrino cross-section sigma_nuN is not measured at energies above 5.2 x 10^13 eV. Although QCD extrapolations offer motivated guesses for sigma_nuN, new physics could intervene to provide a surprise. Therefore, we present the acceptances of horizontal (HAS) and upgoing (UAS) air showers as a function of sigma_nuN over the interesting range 10^(-34) to 10^(-30) cm^2. The dependences of acceptances on neutrino energy, shower-threshold energy, shower length and column density, and cloud layers are also studied. For UAS, we present acceptances for events over land (rock), and over the ocean (water). The latter are larger by about an order of magnitude, thus favoring space-based detectors. We revisit the idea of Kusenko and Weiler to infer sigma_nuN at E_nu ~ 10^20 eV from the ratio of HAS-to-UAS events, and find favorable results. Our calculation is mostly analytic. Included in the UAS calculation are realistic energy-losses for taus, and Earth-curvature effects.

astro-ph↗