Quantifying cardiovascular parameters like ejection fraction in zebrafish as a host of biological investigations has been extensively studied. Since current manual monitoring techniques are time-consuming and fallible, several image processing frameworks have been proposed to automate the process. Most of these works rely on supervised deep-learning architectures. However, supervised methods tend to be overfitted on their training dataset. This means that applying the same framework to new data with different imaging setups and mutant types can severely decrease performance. We have developed a Zebrafish Automatic Cardiovascular Assessment Framework (ZACAF) to quantify the cardiac function in zebrafish. In this work, we further applied data augmentation, Transfer Learning (TL), and Test Time Augmentation (TTA) to ZACAF to improve the performance for the quantification of cardiovascular function quantification in zebrafish. This strategy can be integrated with the available frameworks to aid other researchers. We demonstrate that using TL, even with a constrained dataset, the model can be refined to accommodate a novel microscope setup, encompassing diverse mutant types and accommodating various video recording protocols. Additionally, as users engage in successive rounds of TL, the model is anticipated to undergo substantial enhancements in both generalizability and accuracy. Finally, we applied this approach to assess the cardiovascular function in nrap mutant zebrafish, a model of cardiomyopathy.
Recently, a range of neural network-based methods for image rendering have been introduced. One such widely-researched neural radiance field (NeRF) relies on a neural network to represent 3D scenes, allowing for realistic view synthesis from a small number of 2D images. However, most NeRF models are constrained by long training and inference times. In comparison, Gaussian Splatting (GS) is a novel, state-of-the-art technique for rendering points in a 3D scene by approximating their contribution to image pixels through Gaussian distributions, warranting fast training and swift, real-time rendering. A drawback of GS is the absence of a well-defined approach for its conditioning due to the necessity to condition several hundred thousand Gaussian components. To solve this, we introduce the Gaussian Mesh Splatting (GaMeS) model, which allows modification of Gaussian components in a similar way as meshes. We parameterize each Gaussian component by the vertices of the mesh face. Furthermore, our model needs mesh initialization on input or estimated mesh during training. We also define Gaussian splats solely based on their location on the mesh, allowing for automatic adjustments in position, scale, and rotation during animation. As a result, we obtain a real-time rendering of editable GS.
Current methods to characterize embedded planets in protoplanetary disc observations are severely limited either in their ability to fully account for the observed complex physics or in their computational and time costs. To address this shortcoming, we developed DBNets: a deep learning tool, based on convolutional neural networks, that analyses substructures observed in the dust continuum emission of protoplanetary discs to quickly infer the mass of allegedly embedded planets. We focussed on developing a method to reliably quantify not only the planet mass, but also the associated uncertainty introduced by our modelling and adopted techniques. Our tests gave promising results achieving an 87% reduction of the log Mp mean squared error with respect to an analytical formula fitted on the same data (DBNets metrics: lmse 0.016, r2-score 97%). With the goal of providing the final user of DBNets with all the tools needed to interpret their measurements and decide on their significance, we extensively tested our tool on out-of-distribution data. We found that DBNets can identify inputs strongly outside its training scope returning an uncertainty above a specific threshold and we thus provided a rejection criterion that helps determine the significance of the results obtained. Additionally, we outlined some limitations of our tool: it can be reliably applied only on discs observed with inclinations below approximately 60{\deg}, in the optically thin regime, with a resolution 8 times better than the gap radial location and with a signal-to-noise ratio higher than approximately ten. Finally, we applied DBNets to 33 actual observations of protoplanetary discs measuring the mass of 48 proposed planets and comparing our results with the available literature. We confirmed that most of the observed gaps imply planets in the sub-Jupiter regime. DBNets is publicly available at dbnets.fisica.unimi.it.
Vast literature has compared the recordings of biological neurons in the brain to deep neural networks. The ultimate goal is to interpret deep networks or to better understand and encode biological neural systems. Recently, there has been a debate on whether system identification is possible and how much it can tell us about the brain computation. System identification recognizes whether one model is more valid to represent the brain computation over another. Nonetheless, previous work did not consider the time aspect and how video and dynamics (e.g., motion) modelling in deep networks relate to these biological neural systems within a large-scale comparison. Towards this end, we propose a system identification study focused on comparing single image vs. video understanding models with respect to the visual cortex recordings. Our study encompasses two sets of experiments; a real environment setup and a simulated environment setup. The study also encompasses more than 30 models and, unlike prior works, we focus on convolutional vs. transformer-based, single vs. two-stream, and fully vs. self-supervised video understanding models. The goal is to capture a greater variety of architectures that model dynamics. As such, this signifies the first large-scale study of video understanding models from a neuroscience perspective. Our results in the simulated experiments, show that system identification can be attained to a certain level in differentiating image vs. video understanding models. Moreover, we provide key insights on how video understanding models predict visual cortex responses; showing video understanding better than image understanding models, convolutional models are better in the early-mid regions than transformer based except for multiscale transformers that are still good in predicting these regions, and that two-stream models are better than single stream.
Successful content moderation in online platforms relies on a human-AI collaboration approach. A typical heuristic estimates the expected harmfulness of a post and uses fixed thresholds to decide whether to remove it and whether to send it for human review. This disregards the prediction uncertainty, the time-varying element of human review capacity and post arrivals, and the selective sampling in the dataset (humans only review posts filtered by the admission algorithm). In this paper, we introduce a model to capture the human-AI interplay in content moderation. The algorithm observes contextual information for incoming posts, makes classification and admission decisions, and schedules posts for human review. Only admitted posts receive human reviews on their harmfulness. These reviews help educate the machine-learning algorithms but are delayed due to congestion in the human review system. The classical learning-theoretic way to capture this human-AI interplay is via the framework of learning to defer, where the algorithm has the option to defer a classification task to humans for a fixed cost and immediately receive feedback. Our model contributes to this literature by introducing congestion in the human review system. Moreover, unlike work on online learning with delayed feedback where the delay in the feedback is exogenous to the algorithm's decisions, the delay in our model is endogenous to both the admission and the scheduling decisions. We propose a near-optimal learning algorithm that carefully balances the classification loss from a selectively sampled dataset, the idiosyncratic loss of non-reviewed posts, and the delay loss of having congestion in the human review system. To the best of our knowledge, this is the first result for online learning in contextual queueing systems and hence our analytical framework may be of independent interest.
We present a polynomial-time algorithm for online differentially private synthetic data generation. For a data stream within the hypercube $[0,1]^d$ and an infinite time horizon, we develop an online algorithm that generates a differentially private synthetic dataset at each time $t$. This algorithm achieves a near-optimal accuracy bound of $O(t^{-1/d}\log(t))$ for $d\geq 2$ and $O(t^{-1}\log^{4.5}(t))$ for $d=1$ in the 1-Wasserstein distance. This result generalizes the previous work on the continual release model for counting queries to include Lipschitz queries. Compared to the offline case, where the entire dataset is available at once, our approach requires only an extra polylog factor in the accuracy bound.
Minimum Bayes Risk (MBR) decoding is a text generation technique that has been shown to improve the quality of machine translations, but is expensive, even if a sampling-based approximation is used. Besides requiring a large number of sampled sequences, it requires the pairwise calculation of a utility metric, which has quadratic complexity. In this paper, we propose to approximate pairwise metric scores with scores calculated against aggregated reference representations. This changes the complexity of utility estimation from $O(n^2)$ to $O(n)$, while empirically preserving most of the quality gains of MBR decoding. We release our source code at https://github.com/ZurichNLP/mbr
This study addresses the vital issue of real-time flood detection and management. It innovatively combines advanced deep learning models with Large language models (LLM), enhancing flood monitoring and response capabilities. This approach addresses the limitations of current methods by offering a more accurate, versatile, user-friendly and accessible solution. The integration of UNet, RDN, and ViT models with natural language processing significantly improves flood area detection in diverse environments, including using aerial and satellite imagery. The experimental evaluation demonstrates the models' efficacy in accurately identifying and mapping flood zones, showcasing the project's potential in transforming environmental monitoring and disaster management fields.
Interpreting deep learning time series models is crucial in understanding the model's behavior and learning patterns from raw data for real-time decision-making. However, the complexity inherent in transformer-based time series models poses challenges in explaining the impact of individual features on predictions. In this study, we leverage recent local interpretation methods to interpret state-of-the-art time series models. To use real-world datasets, we collected three years of daily case data for 3,142 US counties. Firstly, we compare six transformer-based models and choose the best prediction model for COVID-19 infection. Using 13 input features from the last two weeks, we can predict the cases for the next two weeks. Secondly, we present an innovative way to evaluate the prediction sensitivity to 8 population age groups over highly dynamic multivariate infection data. Thirdly, we compare our proposed perturbation-based interpretation method with related work, including a total of eight local interpretation methods. Finally, we apply our framework to traffic and electricity datasets, demonstrating that our approach is generic and can be applied to other time-series domains.
This paper compares orthogonal time frequency space (OTFS) modulation and single-carrier frequency division multiple access (SC-FDMA). It shows that these are equivalent except for a set of linear phase shifts, applied to the transmit/receive data symbols, which can be absorbed into the channel. Through mathematical and numerical analysis, it is confirmed that SC-FDMA is in fact a delay-Doppler domain multiplexing technique that can achieve the same performance gains as those of OTFS in time-varying wireless environments. This is a promising result as SC-FDMA is already a part of the current wireless standards. The derivations in this paper also shed light on the time-frequency resources used by the delay-Doppler domain data symbols with the fine granularity of delay and Doppler spacings. While comparing the detection performance of the two waveforms, a timing offset (TO) estimation technique with orders of magnitude higher accuracy than the existing solutions in the literature is proposed. From multiple access viewpoint, the underlying tile structures in the time-frequency domain for OTFS and SC-FDMA are discussed. Finally, multiuser input-output relationships for both waveforms in the uplink are derived.