arXiv subjects

Francisco Pereira

Publications and source records attributed to Francisco Pereira.

At least 19 recordsLinked to original sources

arXiv2026

Interpretable factorization of clinical questionnaires to identify latent factors of psychopathology

Psychiatry research seeks to understand the manifestations of psychopathology in behavior, as measured in questionnaire data, by identifying a small number of latent factors that explain them. While factor analysis is the canonical tool for this purpose, the resulting factors may not be interpretable, and may also be subject to confounding variables. Moreover, missing data are common, and explicit imputation is often required. To overcome these limitations, we introduce Interpretability Constrained Questionnaire Factorization (ICQF), a non-negative matrix factorization method with regularization tailored for questionnaire data. Our method aims to promote factor interpretability and solution stability. We provide an optimization procedure with theoretical convergence guarantees, and an automated procedure to determine latent dimensionality accurately. We validate these procedures using realistic synthetic data. We demonstrate the effectiveness of our method in a widely used general-purpose questionnaire, in two independent datasets (the Healthy Brain Network and Adolescent Brain Cognitive Development studies). Specifically, we show that ICQF preserves diagnostic information across a range of disorders, outperforming competing methods for smaller dataset sizes, and improves interpretability, as assessed by our clinical research collaborators and co-authors. This suggests that the regularization in our method matches domain characteristics, in addition to satisfying qualitative desiderata.

cs.LG↗

arXiv2026

Fast Penalized Generalized Estimating Equations for Large Longitudinal Functional Datasets

Longitudinal binary or count functional data are common in neuroscience, but are often too large to analyze with existing functional regression methods. We propose one-step penalized generalized estimating equations that supports generalized functional outcomes (e.g., count, binary, proportion, continuous-valued) and is fast even when datasets have a large number of clusters and large cluster sizes. The method applies to functional and scalar covariates and the one-step estimation framework enables efficient smoothing parameter selection, bootstrapping, and joint confidence interval construction. Importantly, this semi-parametric approach yields coefficient confidence intervals that are provably valid asymptotically even under working correlation misspecification. By developing a general theory for adaptive one-step M-estimation, we prove that the coefficient estimates are asymptotically normal and as efficient as the fully-iterated estimator; we verify these theoretical properties in simulations. We illustrate the benefits of our approach for analyzing large-scale neural recordings by applying it to a recent calcium imaging dataset published in Nature. We show that our method reveals important timing effects obscured in non-functional analyses. In doing so, we also demonstrate scaling to common neuroscience dataset sizes: the one-step estimator fits to a dataset with 150,000 (binary) functional outcomes, each observed at 120 functional domain points, in only 6.5 minutes on a laptop without parallelization. We release our methods in the R package 'fastfGEE', which supports a wide range of link functions and working covariances.

stat.ME↗

arXiv2026

Similarity-based representation factorization for revealing interpretable dimensions in representational data

The study of representations is widespread across fields, including neuroscience, psychology, and artificial intelligence. While representations are often studied and compared through similarities between stimuli, current methods provide only limited access to the dimensions that shape these representations and are often limited in interpretability. To overcome these challenges, here we introduce Similarity-Based Representation Factorization (SRF), a general computational method for recovering low-dimensional, non-negative, interpretable embeddings from similarity matrices derived from measured data. Across simulations and many neural, behavioral, and computational datasets, SRF recovers interpretable dimensions from diverse forms of representational data, even for very sparsely sampled, incomplete data. The dimensions derived from these datasets match those obtained by task-specific models, predict independent behavioral properties, improve exploratory analysis, and offer higher power for confirmatory hypothesis testing than comparing similarity matrices. Together, these results establish SRF as a general-purpose method with broad applications for uncovering, understanding, and using the dimensions underlying representations.

cs.CV↗

arXiv2026

Long-range deformations in Gaussian States

Imaginary-time evolution by a local Hamiltonian cannot induce a phase transition in one dimension, but longer-range interactions may subvert such constraints. Starting from the ground state of the Kitaev Majorana chain, we modify the wave function by an imaginary-time evolution generated by a quadratic Hamiltonian with power-law couplings that enhance pairing correlations, typically of the form $1/r^α$, where $r$ is the distance between two sites. As the state remains Gaussian, entanglement and correlation functions can be computed analytically. We find that the decay exponent $α$ controls three distinct infrared regimes: for $α>1$, the deformation produces only smooth crossovers at finite deformation strength, while the topological regime is reached only asymptotically as the deformation strength tends to infinity. At $α=1$, the deformation induces an immediate flow to the topological phase: an infinitesimal deformation strength drives the system to a topological regime, and in a particular case, an emergent Kramers-Wannier symmetry enforces Ising-like scaling at long distances. For $α<1$, the deformed state shows the same critical-like behavior for all non-zero deformation strength. We observe that even with arbitrarily long-range interactions, these models do not display a sharp phase transition at non-zero deformation strength.

cond-mat.stat-mech↗

arXiv2026

Characterizing Universal Object Representations Across Vision Models

Deep neural networks trained with different architectures, objectives, and datasets have been reported to converge on similar visual representations. However, what remains unknown is which visual properties models actually converge on and which factors may underlie this convergence. To address this, we decompose the object similarity structure of 162 diverse vision models into a small set of non-negative dimensions. To determine universal versus model-specific dimensions, we then estimate how often each dimension reappears across models. In contrast to model-specific dimensions, universal dimensions are more interpretable and more strongly driven by conceptual image properties, indicating the relevance of interpretability and semantic content as implicit factors driving universality across models. Differences in architecture, objective function, training data, model size, and model performance do not explain the emergence of universal dimensions. However, models with more universal dimensions also better predict macaque IT activity and human similarity judgments, suggesting that universality reflects representations relevant to biological vision. These findings have important implications for understanding the emergent representations underlying deep neural network models and their alignment with biological vision.

cs.CV↗

arXiv2026

Enhancing Infant Crying Detection with Gradient Boosting for Improved Emotional and Mental Health Diagnostics

Infant crying can serve as a crucial indicator of various physiological and emotional states. This paper introduces a comprehensive approach detecting infant cries within audio data. We integrate Wav2Vec with traditional audio features and employ Gradient Boosting Machines for cry classification. We validate our approach on a real world dataset, demonstrating significant performance improvements over existing methods.

eess.AS↗

arXiv2025

Nonparametric causal inference for optogenetics: sequential excursion effects for dynamic regimes

Optogenetics is a powerful neuroscience technique for studying how neural circuit manipulation affects behavior. Standard analysis conventions discard information and severely limit the scope of the causal questions that can be probed. To address this gap, we 1) draw connections to the causal inference literature on sequentially randomized experiments, 2) propose a non-parametric framework for analyzing "open-loop" (static regime) optogenetics behavioral experiments, 3) derive extensions of history-restricted marginal structural models for dynamic treatment regimes with positivity violations for "closed-loop" designs, and 4) propose a taxonomy of identifiable causal effects that encompass a far richer collection of scientific questions compared to standard methods. From another view, our work extends "excursion effect" methods, popularized recently in the mobile health literature, to enable estimation of causal contrasts for treatment sequences in the presence of positivity violations. We describe sufficient conditions for identifiability of the proposed causal estimands, and provide asymptotic statistical guarantees for a proposed inverse probability-weighted estimator, a multiply-robust estimator (for two intervention timepoints), a framework for hypothesis testing, and a computationally scalable implementation. Finally, we apply our framework to data from a recent neuroscience study and show how it provides insight into causal effects of optogenetics on behavior that are obscured by standard analyses.

stat.ME↗

arXiv2025

More Experts Than Galaxies: Conditionally-overlapping Experts With Biologically-Inspired Fixed Routing

The evolution of biological neural systems has led to both modularity and sparse coding, which enables energy efficiency and robustness across the diversity of tasks in the lifespan. In contrast, standard neural networks rely on dense, non-specialized architectures, where all model parameters are simultaneously updated to learn multiple tasks, leading to interference. Current sparse neural network approaches aim to alleviate this issue but are hindered by limitations such as 1) trainable gating functions that cause representation collapse, 2) disjoint experts that result in redundant computation and slow learning, and 3) reliance on explicit input or task IDs that limit flexibility and scalability. In this paper we propose Conditionally Overlapping Mixture of ExperTs (COMET), a general deep learning method that addresses these challenges by inducing a modular, sparse architecture with an exponential number of overlapping experts. COMET replaces the trainable gating function used in Sparse Mixture of Experts with a fixed, biologically inspired random projection applied to individual input representations. This design causes the degree of expert overlap to depend on input similarity, so that similar inputs tend to share more parameters. This results in faster learning per update step and improved out-of-sample generalization. We demonstrate the effectiveness of COMET on a range of tasks, including image classification, language modeling, and regression, using several popular deep learning architectures.

cs.LG↗

arXiv2024

Reliability and predictability of phenotype information from functional connectivity in large imaging datasets

One of the central objectives of contemporary neuroimaging research is to create predictive models that can disentangle the connection between patterns of functional connectivity across the entire brain and various behavioral traits. Previous studies have shown that models trained to predict behavioral features from the individual's functional connectivity have modest to poor performance. In this study, we trained models that predict observable individual traits (phenotypes) and their corresponding singular value decomposition (SVD) representations - herein referred to as latent phenotypes from resting state functional connectivity. For this task, we predicted phenotypes in two large neuroimaging datasets: the Human Connectome Project (HCP) and the Philadelphia Neurodevelopmental Cohort (PNC). We illustrate the importance of regressing out confounds, which could significantly influence phenotype prediction. Our findings reveal that both phenotypes and their corresponding latent phenotypes yield similar predictive performance. Interestingly, only the first five latent phenotypes were reliably identified, and using just these reliable phenotypes for predicting phenotypes yielded a similar performance to using all latent phenotypes. This suggests that the predictable information is present in the first latent phenotypes, allowing the remainder to be filtered out without any harm in performance. This study sheds light on the intricate relationship between functional connectivity and the predictability and reliability of phenotypic information, with potential implications for enhancing predictive modeling in the realm of neuroimaging research.

q-bio.NC↗

arXiv2024

Peregrine: ML-based Malicious Traffic Detection for Terabit Networks

Malicious traffic detectors leveraging machine learning (ML), namely those incorporating deep learning techniques, exhibit impressive detection capabilities across multiple attacks. However, their effectiveness becomes compromised when deployed in networks handling Terabit-speed traffic. In practice, these systems require substantial traffic sampling to reconcile the high data plane packet rates with the comparatively slower processing speeds of ML detection. As sampling significantly reduces traffic observability, it fundamentally undermines their detection capability. We present Peregrine, an ML-based malicious traffic detector for Terabit networks. The key idea is to run the detection process partially in the network data plane. Specifically, we offload the detector's ML feature computation to a commodity switch. The Peregrine switch processes a diversity of features per-packet, at Tbps line rates - three orders of magnitude higher than the fastest detector - to feed the ML-based component in the control plane. Our offloading approach presents a distinct advantage. While, in practice, current systems sample raw traffic, in Peregrine sampling occurs after feature computation. This essential trait enables computing features over all traffic, significantly enhancing detection performance. The Peregrine detector is not only effective for Terabit networks, but it is also energy- and cost-efficient. Further, by shifting a compute-heavy component to the switch, it saves precious CPU cycles and improves detection throughput.

cs.NI↗

arXiv2023

Automatic Parallelization of Software Network Functions

Software network functions (NFs) trade-off flexibility and ease of deployment for an increased challenge of performance. The traditional way to increase NF performance is by distributing traffic to multiple CPU cores, but this poses a significant challenge: how to parallelize an NF without breaking its semantics? We propose Maestro, a tool that analyzes a sequential implementation of an NF and automatically generates an enhanced parallel version that carefully configures the NIC's Receive Side Scaling mechanism to distribute traffic across cores, while preserving semantics. When possible, Maestro orchestrates a shared-nothing architecture, with each core operating independently without shared memory coordination, maximizing performance. Otherwise, Maestro choreographs a fine-grained read-write locking mechanism that optimizes operation for typical Internet traffic. We parallelized 8 software NFs and show that they generally scale-up linearly until bottlenecked by PCIe when using small packets or by 100Gbps line-rate with typical Internet traffic. Maestro further outperforms modern hardware-based transactional memory mechanisms, even for challenging parallel-unfriendly workloads.

cs.NI↗

arXiv2022

Testing for context-dependent changes in neural encoding in naturalistic experiments

We propose a decoding-based approach to detect context effects on neural codes in longitudinal neural recording data. The approach is agnostic to how information is encoded in neural activity, and can control for a variety of possible confounding factors present in the data. We demonstrate our approach by determining whether it is possible to decode location encoding from prefrontal cortex in the mouse and, further, testing whether the encoding changes due to task engagement.

stat.ML↗

arXiv2022

VICE: Variational Interpretable Concept Embeddings

A central goal in the cognitive sciences is the development of numerical models for mental representations of object concepts. This paper introduces Variational Interpretable Concept Embeddings (VICE), an approximate Bayesian method for embedding object concepts in a vector space using data collected from humans in a triplet odd-one-out task. VICE uses variational inference to obtain sparse, non-negative representations of object concepts with uncertainty estimates for the embedding values. These estimates are used to automatically select the dimensions that best explain the data. We derive a PAC learning bound for VICE that can be used to estimate generalization performance or determine a sufficient sample size for experimental design. VICE rivals or outperforms its predecessor, SPoSE, at predicting human behavior in the triplet odd-one-out task. Furthermore, VICE's object representations are more reproducible and consistent across random initializations, highlighting the unique advantage of using VICE for deriving interpretable embeddings from human behavior.

cs.LG↗

arXiv2022

Representation learning of rare temporal conditions for travel time prediction

Predicting travel time under rare temporal conditions (e.g., public holidays, school vacation period, etc.) constitutes a challenge due to the limitation of historical data. If at all available, historical data often form a heterogeneous time series due to high probability of other changes over long periods of time (e.g., road works, introduced traffic calming initiatives, etc.). This is especially prominent in cities and suburban areas. We present a vector-space model for encoding rare temporal conditions, that allows coherent representation learning across different temporal conditions. We show increased performance for travel time prediction over different baselines when utilizing the vector-space encoding for representing the temporal setting.

stat.ML↗

arXiv2021

Mental representations of objects reflect the ways in which we interact with them

In order to interact with objects in our environment, humans rely on an understanding of the actions that can be performed on them, as well as their properties. When considering concrete motor actions, this knowledge has been called the object affordance. Can this notion be generalized to any type of interaction that one can have with an object? In this paper we introduce a method to represent objects in a space where each dimension corresponds to a broad mode of interaction, based on verb selectional preferences in text corpora. This object embedding makes it possible to predict human judgments of verb applicability to objects better than a variety of alternative approaches. Furthermore, we show that the dimensions in this space can be used to predict categorical and functional dimensions in a state-of-the-art mental representation of objects, derived solely from human judgements of object similarity. These results suggest that interaction knowledge accounts for a large part of mental representations of objects.

cs.CL↗

arXiv2020

Improving the Interpretability of fMRI Decoding using Deep Neural Networks and Adversarial Robustness

Deep neural networks (DNNs) are being increasingly used to make predictions from functional magnetic resonance imaging (fMRI) data. However, they are widely seen as uninterpretable "black boxes", as it can be difficult to discover what input information is used by the DNN in the process, something important in both cognitive neuroscience and clinical applications. A saliency map is a common approach for producing interpretable visualizations of the relative importance of input features for a prediction. However, methods for creating maps often fail due to DNNs being sensitive to input noise, or by focusing too much on the input and too little on the model. It is also challenging to evaluate how well saliency maps correspond to the truly relevant input information, as ground truth is not always available. In this paper, we review a variety of methods for producing gradient-based saliency maps, and present a new adversarial training method we developed to make DNNs robust to input noise, with the goal of improving interpretability. We introduce two quantitative evaluation procedures for saliency map methods in fMRI, applicable whenever a DNN or linear model is being trained to decode some information from imaging data. We evaluate the procedures using a synthetic dataset where the complex activation structure is known, and on saliency maps produced for DNN and linear models for task decoding in the Human Connectome Project (HCP) dataset. Our key finding is that saliency maps produced with different methods vary widely in interpretability, in both in synthetic and HCP fMRI data. Strikingly, even when DNN and linear models decode at comparable levels of performance, DNN saliency maps score higher on interpretability than linear model saliency maps (derived via weights or gradient). Finally, saliency maps produced with our adversarial training method outperform those from other methods.

cs.LG↗

arXiv2020

A Deep Neural Network Tool for Automatic Segmentation of Human Body Parts in Natural Scenes

This short article describes a deep neural network trained to perform automatic segmentation of human body parts in natural scenes. More specifically, we trained a Bayesian SegNet with concrete dropout on the Pascal-Parts dataset to predict whether each pixel in a given frame was part of a person's hair, head, ear, eyebrows, legs, arms, mouth, neck, nose, or torso.

cs.CV↗

arXiv2019

Knowing what you know in brain segmentation using Bayesian deep neural networks

In this paper, we describe a Bayesian deep neural network (DNN) for predicting FreeSurfer segmentations of structural MRI volumes, in minutes rather than hours. The network was trained and evaluated on a large dataset (n = 11,480), obtained by combining data from more than a hundred different sites, and also evaluated on another completely held-out dataset (n = 418). The network was trained using a novel spike-and-slab dropout-based variational inference approach. We show that, on these datasets, the proposed Bayesian DNN outperforms previously proposed methods, in terms of the similarity between the segmentation predictions and the FreeSurfer labels, and the usefulness of the estimate uncertainty of these predictions. In particular, we demonstrated that the prediction uncertainty of this network at each voxel is a good indicator of whether the network has made an error and that the uncertainty across the whole brain can predict the manual quality control ratings of a scan. The proposed Bayesian DNN method should be applicable to any new network architecture for addressing the segmentation problem.

cs.CV↗