KU Leuven Research Data Repository
Not a member yet
662 research outputs found
Sort by
LARMix: Latency-Aware Routing in Mix Networks
No full textThe LARMix repository is designed to replicate the results (including figures and tables) from the paper titled "LARMix: Latency-Aware Routing in Mix Networks." Users also have the option to customize parameters to their preferences. The results are stored in two formats: one as a .json file and the other as figures and tables within directories named Figure/ and Table/
Bootstrapping for BGV and BFV Revisited
No full textBootstrapping is an algorithm in homomorphic encryption that reduces ciphertext noise. The code implements bootstrapping for the BGV and GBFV schemes in Magma. It also allows to export GBFV functionality to the SEAL library by means of high-level instruction traces. The code is subdivided in several folders that are responsible for arithmetic, cryptographic and bootstrapping-related operations
COherent Sampling ring Oscillator (COSO) based True Random Number Generator (TRNG)
No full textThis dataset provides a reference implementation of the COherent Sampling ring Oscillator (COSO) based True Random Number Generator (TRNG), leveraging configurable Ring Oscillators (ROs). It includes raw measurement data to support the flexibility and portability claims associated with the TRNG design. Figure generation scripts are provided to visualize the raw measurement data, and stochastic model simulation scripts are included to explore the entropy-throughput tradeoff
Replication Data for: Self-Learning for Personalized Keyword Spotting on Ultra-Low-Power Audio Sensors
No full textThis repository includes the data for reproducing the results of the paper: "Self-Learning for Personalized Keyword Spotting on Ultra-Low-Power Audio Sensors". In this work, we recorded a dataset of speech samples with our microphone sensor after replaying a subset of data from the HeySnips dataset using a speaker.
The collected speech data (total of 400 samples) are split between a testset and a trainset, both including "Hey Snips" utterances and non-"Hey Snips" utterances.
In particular, the data of the testset is composed by recordings from 20 random speakers from the original testset.
After the recording, the data were fed to our DNN models deployed on devices.
Initially, a per-speaker prototype vector is computed by feeding three audio recordings of the target keywords.
Next, the audio tracks of the training set are processed with a sliding window approach to compute the distance with respect to the prototype and assign pseudo-labels for the self-learning task.
The dataset is, therefore, composed of two main partitions. First, the "recorded_speech_data" includes the audio recordings. Note that this dataset is under restricted access to not violate the terms of access of the original dataset. Second, the "processed_outputs" includes the output of the processing, i.e. the measured distances.
By using the dataset in combination with the associated code, every user will be able to reproduce the results of the paper
MSNER: Multilingual Spoken Named Entity Recognition Dataset
No full textWhile extensively explored in text-based tasks, Named Entity Recognition (NER) remains largely neglected in spoken language understanding. Existing resources are limited to a single, English-only dataset. This paper addresses this gap by introducing MSNER, a freely available, multilingual speech corpus annotated with named entities. It provides annotations to the VoxPopuli dataset in four languages (Dutch, French, German, and Spanish). We have also releasing an efficient annotation tool that leverages automatic pre-annotations for faster manual refinement. This results in 590 and 15 hours of silver-annotated speech for training and validation, alongside a 17-hour, manually-annotated evaluation set. We further provide an analysis comparing silver and gold annotations. Finally, we present baseline NER models to stimulate further research on this newly available dataset
Replication Data for: Denoising of Photogrammetric Dummy Head Ear Point Clouds for Individual Head-Related Transfer Functions Computation
No full textThis repository contains data and code for point cloud denoising of photogrammetric dummy head ear scans used in computing Head-Related Transfer Functions (HRTFs).
Included in this repository are the point cloud data for training and testing the denoising models, along with the learned parameters of the trained models. Additionally, the numerically simulated HRTFs of the original and denoised scans of the KU100 dummy head geometry are provided as time domain Head-Related Impulse Responses (HRIRs). Test code is provided to facilitate the processing and visualisation of the data.
The purpose of the data and code in this repository is to analyse the effect of ear point cloud denoising on the HRTFs computed on the scanned and processed geometries. This analysis includes evaluating geometrical metrics computed on the initial and denoised scans, as well as objective and perceptually inspired HRTF metrics on the corresponding HRIRs.
For further details, please refer to the "readme.md" file.</p
Climate change communication on the websites of Flemish political parties
No full textThe data in this repository offers a sample of online climate change communication by political parties in Flanders, the northern, Dutch-speaking part of Belgium. The dataset consists of the party positions ("standpunten" in Dutch) that the seven main political parties in Flanders posted on their websites, more specifically those that were indexed by the parties as related to the climate, climate change and/or the environment.
The seven parties included in this dataset are those that are represented, at the time of the data collection, in the Flemish parliament. These are: N-VA (New-Flemish Alliance), Vlaams Belang (Flemish Interest), CD&V (Christian Democratic & Flemish), Open VLD (Open Flemish Liberals and Democrats), Vooruit (Forward), Groen (Green), and PVDA (Labour Party of Belgium).
The choice to include only those data that were explicitly indicated by the political parties as pertaining to climate change and/or the environment was deliberate, as it allows for a targeted examination of how political parties communicate about climate-related and environmental issues and possible solutions. It also means that the topics covered are more expansive for some parties than for others. For instance, while the green party ‘Groen’ takes a comprehensive view of climate-related policies (which are presented as also including energy, mobility, food and agriculture, and sustainable economy), the regionalist party ‘N-VA’ has opted for a more diversified approach, with environmental and climate-related policies presented as a separate category from energy, mobility, agriculture, etc. Despite its limitations, this procedure ensured that the data of political climate change communication is, indeed, explicitly understood by the parties themselves as relating to climate change
Replication Data for: Unlocking Temperature Reduction of Cogeneration District Heating Networks through Automated Substation Retrofit
No full textA dataset containing the structure, input parameters and optimization results of the heating network and substations used in the test case of the paper "Unlocking Temperature Reduction of Cogeneration District Heating Networks through Automated Substation Retrofit". The optimization results can be replicated using the methodology and formulations described in the paper
Trajectory files for the heliospheric model Icarus
No full textThis dataset contains files describing trajectories of planets in the solar system and several satellites needed for running the heliospheric model Icarus. The files allow the model to track the positions of the objects during the simulation and to save data specifically for the object locations at each point in time. The files can be downloaded using a Python script (provided as part of Icarus) which downloades only the files necessary for given time
Instance files for PnP systems
No full textThis data repository contains the problem instance files that were used to validate the proposed optimization approaches for the multi-robot TD-PnP-TW problem. The instances vary in problem scale, were both small-scale (10.000 objects). Moreover, they vary in the object density that is for the objects on the conveyor belt. With these both one-shot setting and continuous simulations can be performed to analyze and validate the optimization approaches