The 2020 Census Demographic and Housing Characteristics Noisy Measurement File is an intermediate output of the 2020 Census Disclosure Avoidance System (DAS) TopDown Algorithm (TDA) (as described in Abowd, J. et al [2022], and implemented in DAS_2020_DHC_Production_Code/das_decennial/programs/engine/primitives.py at main uscensusbureau/DAS_2020_DHC_Production_Code (github.com) The 2020 Census Demographic and Housing Characteristics Noisy Measurement File includes zero-Concentrated Differentially Private (zCDP) (Bun, M. and Steinke, T [2016]) noisy measurements, implemented via the discrete Gaussian mechanism (Cannone C., et al., [2023] ), which added positive or negative integer-valued noise to each of the resulting counts. These are estimated counts of individuals and housing units included in the 2020 Census Edited File (CEF), which includes confidential data collected in the 2020 Census of Population and Housing. The noisy measurements included in this file were subsequently post-processed by the TopDown Algorithm (TDA) to produce the Census Demographic and Housing Characteristics Summary File. In addition to the noisy measurements, constraints based on invariant calculations --- counts computed without noise --- are also included (with the exception of the state-level total populations, which can be sourced separately from data.census.gov). The Noisy Measurement File was produced using the official "production settings," the final set of algorithmic parameters and privacy-loss budget allocations that were used to produce the 2020 Census Redistricting Data (P.L. 94-171) Summary File and the 2020 Census Demographic and Housing Characteristics File. The noisy measurements are produced in an early stage of the TDA. Afterward, these noisy measurements are post-processed to ensure internal and hierarchical consistency within the resulting tables. The Census Bureau has released these noisy measurements to enable data users to evaluate the impact of disclosure avoidance variability on 2020 Census data. The 2020 Census Demographic and Housing Characteristics (DHC) Noisy Measurement File has been cleared for public dissemination by the Census Bureau Disclosure Review Board (CBDRB-FY22-DSEP-004). These data are available for download (i.e. not restricted access). Due to their size, they must be downloaded through the link on this metadata page and not through the standard ICPSR download. The link will take you to the Globus site where these data are housed. A README file is located in the Globus repository. Please refer to that for pertinent information. The Globus holding site requires users to create an account to access these data. Accounts can be created through existing institutional access and by personal access. Please see the Globus "How to get Started" page for more information.

These are the time-averaged noisy datasets for both Newtonian and non-Newtonian simulations

This dataset is based on MS COCO that have 20% of data randomly shuffled to simulate noisy correspondence.

Replication files for: Dahlhaus, T. and Gambetti L., Noisy Monetary Policy Announcements, Journal of Applied Econometrics.

MainProg_JAE.m replicates the analysis in Section 3.2-3.5. Specifically, it loads the data, estimates the news and noise shocks and responses, and estimates the responses to Delphic and Odyssean shocks based on noise-free data. Run NoiseNewsShock.m to produce Figures 1,2, and 3. Run InfoShock1.m and InfoShock2.m to produce Figures 4 and 5, respectively. Main_FinancialVars_monthly.m and Main_FinancialVars_HF replicate the analysis of Section 3.6 loading the data and producing the variance decomposition of news and noise shocks for monthly aggregates of financial variables and the high-frequency change of those variables around FOMC dates, respectively. Financial_responses.m produces and plots the responses of financial variables (Figure 6 and Figure 7).

The 2010 Census Production Settings Redistricting Data (P.L. 94-171) Demonstration Noisy Measurement Files are an intermediate output of the 2020 Census Disclosure Avoidance System (DAS) TopDown Algorithm (TDA) (as described in Abowd, J. et al [2022], and implemented in https://github.com/uscensusbureau/DAS_2020_Redistricting_Production_Code). The NMF was produced using the official "production settings," the final set of algorithmic parameters and privacy-loss budget allocations that were used to produce the 2020 Census Redistricting Data (P.L. 94-171) Summary File and the 2020 Census Demographic and Housing Characteristics File. The NMF consists of the full set of privacy-protected statistical queries (counts of individuals or housing units with particular combinations of characteristics) of confidential 2010 Census data relating to the redistricting data portion of the 2010 Demonstration Data Products Suite - Redistricting and Demographic and Housing Characteristics File - Production Settings (2023-04-03). These statistical queries, called "noisy measurements" were produced under the zero-Concentrated Differential Privacy framework (Bun, M. and Steinke, T [2016]; see also Dwork C. and Roth, A. [2014]) implemented via the discrete Gaussian mechanism (Cannone C., et al., [2023]), which added positive or negative integer-valued noise to each of the resulting counts. The noisy measurements are an intermediate stage of the TDA prior to the post-processing the TDA then performs to ensure internal and hierarchical consistency within the resulting tables. The Census Bureau has released these 2010 Census demonstration data to enable data users to evaluate the expected impact of disclosure avoidance variability on 2020 Census data. The 2010 Census Production Settings Redistricting Data (P.L. 94-171) Demonstration Noisy Measurement Files (2023-04-03) have been cleared for public dissemination by the Census Bureau Disclosure Review Board (CBDRB-FY22-DSEP-004). The data include zero-Concentrated Differentially Private (zCDP) (Bun, M. and Steinke, T [2016]) noisy measurements, implemented via the discrete Gaussian mechanism. These are estimated counts of individuals and housing units included in the 2010 Census Edited File (CEF), which includes confidential data initially collected in the 2010 Census of Population and Housing. The noisy measurements included in this file were subsequently post-processed by the TopDown Algorithm (TDA) to produce the 2010 Census Production Settings Privacy-Protected Microdata File - Redistricting (P.L. 94-171) and Demographic and Housing Characteristics File (2023-04-03) (https://www2.census.gov/programs-surveys/decennial/2020/program-management/data-product- planning/2010-demonstration-data-products/04 Demonstration_Data_Products_Suite/2023-04-03/). As these 2010 Census demonstration data are intended to support study of the design and expected impacts of the 2020 Disclosure Avoidance System, the 2010 CEF records were pre-processed before application of the zCDP framework. This pre-processing converted the 2010 CEF records into the input-file format, response codes, and tabulation categories used for the 2020 Census, which differ in substantive ways from the format, response codes, and tabulation categories originally used for the 2010 Census. The NMF provides estimates of counts of persons in the CEF by various characteristics and combinations of characteristics, including their reported race and ethnicity, whether they were of voting age, whether they resided in a housing unit or one of 7 group quarters types, and their census block of residence, after the addition of discrete Gaussian noise (with the scale parameter determined by the privacy-loss budget allocation for that particular query under zCDP). Noisy measurements of the counts of occupied and vacant housing units by census block are also included. Lastly, data on constraints--information into which no noise was infused by the Disclosure Avoidance System (DAS) and used by the TDA to post-process the noisy measurements into the 2010 Census Production Settings Privacy-Protected Microdata File - Redistricting (P.L. 94-171) and Demographic and Housing Characteristics File (2023-04-03) --are provided. These data are available for download (i.e. not restricted access). Due to their size, they must be downloaded through the link on this

Noisy Data

## Overview

Noisy Data is a dataset for object detection tasks - it contains Tumor annotations for 251 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

The FSDnoisy18k dataset is an open dataset containing 42.5 hours of audio across 20 sound event classes, including a small amount of manually-labeled data and a larger quantity of real-world noisy data. The audio content is taken from Freesound, and the dataset was curated using the Freesound Annotator. The noisy set of FSDnoisy18k consists of 15,813 audio clips (38.8h), and the test set consists of 947 audio clips (1.4h) with correct labels. The dataset features two main types of label noise: in-vocabulary (IV) and out-of-vocabulary (OOV). IV applies when, given an observed label that is incorrect or incomplete, the true or missing label is part of the target class set. Analogously, OOV means that the true or missing label is not covered by those 20 classes.

Deep generative models have demonstrated an excellent ability to generate data by learning their distribution. Despite their unsupervised nature, these models can be implemented in semi-supervised learning scenarios by treating the class labels as additional latent variables. In this article, we propose a deep generative model for semi-supervised learning that offsets label noise, which is a ubiquitous feature in large-scale datasets owing to the high cost of annotation. We assume that noisy labels are generated from true labels and employ a noise transition matrix to describe the transition. We estimate this matrix by adjusting its entries to minimize its difference with the true transition matrix and use the estimated matrix to formulate the objective function for inference, which consists of an evidence lower bound and a classification risk. However, because directly minimizing the latter with noisy labels may result in an inaccurate classifier, we propose a statistically consistent estimator for computing the classification risk solely with noisy data. Empirical results on benchmark datasets demonstrate that the proposed model improves the classification performance over that of the baseline algorithms. We also present a case study on semiconductor manufacturing. Additionally, we empirically show that the proposed model, as a generative model, is capable of reconstructing data even with noisy labels.

Simulation data of an ultralight aircraft.

Data *.mat collects in MAT files all flight data and ASSE coefficients for three manoeuvres:

stall: from 10 s to 40 s

AoS sweep: from 80 s to 110 s

3211 elevator: from 5 s to 40 s

Data *_noise.mat are the same with uncertainty levels described in https://www.mdpi.com/1436482

Dataset

This dataset was created by Shusrith

Released under Apache 2.0

Contents

This data release complements Murray et al. (2023) which presents a framework for incorporating earthquake magnitude estimates based on real-time Global Navigation Satellite System (GNSS) data into the ShakeAlert® earthquake early warning system for the west coast of the United States. Murray et al. (2023) assess the impact of time-dependent noise in GNSS real-time position estimates on the reliability of earthquake magnitudes estimated using such data. To do so they derived peak ground displacement (PGD) estimates from time series of background noise in GNSS real-time positions. These noise-only PGD measurements were used as input to a published empirical relationship to compute magnitude for hypothetical earthquakes that are each defined by an epicentral location and origin time. The data files provided here give the locations of GNSS stations used in the study, the hypothetical epicenters and origin times, and the PGD for each GNSS station for four time windows following each hypothetical origin time. We also provide the epicenters and origin times used to simulate the impact of noisy PGD data in terms of the annual number of spuriously large magnitude estimates that would be generated in the geographic region spanned by California, Oregon, and Washington, United States, due to noise alone. Finally, we include the estimated magnitudes for the annual simulations along with the number of GNSS stations for which the measured PGD exceeding a threshold value that was defined empirically to eliminate unreliable magnitude estimates.

This dataset provides information about the number of properties, residents, and average property values for Noisy Hole Road cross streets in Mashpee, MA.

Noise of Web (NoW) is a challenging noisy correspondence learning (NCL) benchmark for robust image-text matching/retrieval models. It contains 100K image-text pairs consisting of website pages and multilingual website meta-descriptions (98,000 pairs for training, 1,000 for validation, and 1,000 for testing). NoW has two main characteristics: without human annotations and the noisy pairs are naturally captured. The source image data of NoW is obtained by taking screenshots when accessing web pages on mobile user interface (MUI) with 720 $\times$ 1280 resolution, and we parse the meta-description field in the HTML source code as the captions. In NCR (predecessor of NCL), each image in all datasets were preprocessed using Faster-RCNN detector provided by Bottom-up Attention Model to generate 36 region proposals, and each proposal was encoded as a 2048-dimensional feature. Thus, following NCR, we release our the features instead of raw images for fair comparison. However, we can not just use detection methods like Faster-RCNN to extract image features since it is trained on real-world animals and objects on MS-COCO. To tackle this, we adapt APT as the detection model since it is trained on MUI data. Then, we capture the 768-dimensional features of top 36 objects for one image. Due to the automated and non-human curated data collection process, the noise in NoW is highly authentic and intrinsic. The estimated noise ratio of this dataset is nearly 70%.

This article studies costly signaling. The signaling effort is chosen in multiple periods and observed with noise. The signaler benefits from the belief of the market, not directly from the effort or the signal. Optimal signaling behavior in time-varying environments trades off effort-smoothing and influencing belief exactly when it yields a return. If the return to signaling first increases over time and then decreases, then the optimal effort rises slowly, reaches its maximum before the return does, and declines quickly. Advertising data displays this pattern.

Buffer of sound footprints around motorways, RN, RD, communal lanes, railways and infrastructure projects. The data is derived from the sound classification base.

In each department, the prefect shall, by order, identify and classify land transport infrastructure, after having taken the opinion of the municipalities concerned. These data are then integrated into the planning documents in order to enable the noise control approach to take on a preventive arm. Thus, when a construction is planned in an area affected by the noise reported in the Local Urbanisation Plan (PLU), the manufacturer must comply with certain standards in terms of acoustic insulation of the facade.

The sound classification must be reviewed and revised every 5 years. The sound classification applies to all construction authorities (state, department, community of municipalities and municipalities) but also to the railway network and tram lines. — Roadways with more than 5000 vehicles per day. — Intercity railway tracks of more than 50 trains per day. — Urban railway tracks of more than 100 trains per day. — Public transport lanes on a clean site of more than 100 buses or trains per day. — The planned infrastructure (to be taken into account as soon as the act of opening a public inquiry is published, registration in a reserved location in the PLU or the establishment of a project of general interest).

Source: DDTM62/SSERBC/NOISE

This dataset provides information about the number of properties, residents, and average property values for Noisy Branch cross streets in Keaton, KY.

Description: Contains a denoising training and test dataset for deep learning applications.

Training dataset: 20 paired matching noisy and high signal to noise images

Test dataset: 5 paired matching noisy and high signal to noise images

Microscopy data type: Fluorescence microscopy (SiR-DNA) images.

Microscope: Spinning disk confocal microscope with a 20x 0.8 NA objective

Cell type: DCIS.COM Lifeact-RFP cells

File format: .tif (16-bit for fluorescence)

Image size: 1024x1024 (Pixel size: 634 nm)

Normalised dissimilarities between MPs generated from clean data and from data with various noise parameter settings, for each data set.

QUT Research Data Respository Dataset Resource available for download

sample duplications