DaRUS (University of Stuttgart)
Not a member yet
2037 research outputs found
Sort by
Replication Data for: Advanced System Integration: Analyzing OpenAPI Chunking for Retrieval-Augmented Generation
No full textIntegrating multiple (sub-)systems is essential to create advanced Information Systems (ISs). Difficulties mainly arise when integrating dynamic environments across the IS lifecycle, e.g., services not yet existent at design time. A traditional approach is a registry that provides the API documentation of the systems’ endpoints. Large Language Models (LLMs) have shown to be capable of automatically creating system integrations (e.g., as service composition) based on this documentation but require concise input due to input token limitations, especially regarding comprehensive API descriptions. Currently, it is unknown how best to preprocess these API descriptions. Within this work, we (i) analyze the usage of Retrieval Augmented Generation (RAG) for endpoint discovery and the chunking, i.e., preprocessing, of state-of-practice OpenAPIs to reduce the input token length while preserving the most relevant information. To further reduce the input token length for the composition prompt and improve endpoint retrieval, we propose (ii) a Discovery Agent that only receives a summary of the most relevant endpoints and retrieves specification details on demand. We evaluate RAG for endpoint discovery using the RestBench benchmark, first, for the different chunking possibilities and parameters measuring the endpoint retrieval recall, precision, and F1 score. Then, we assess the Discovery Agent using the same test set. With our prototype, we demonstrate how to successfully employ RAG for endpoint discovery to reduce token count. While revealing high values for recall, precision, and F1, further research is necessary to retrieve all requisite endpoints. Our experiments show that for preprocessing, LLM-based and format-specific approaches outperform naïve chunking methods. Relying on an agent further enhances these results as the agent splits the tasks into multiple fine granular subtasks, improving the overall RAG performance in the token count, precision, and F1 score.
Content:
code.zip:Python source code to perform the experiments.
evaluate.py: Script to execute the experiments (Uncomment lines to select the embedding model).
socrag/*: Source code for the RAG.
benchmark/*: RestBench specification.
results.zip:Results of the RAG experiments (in the folder /results/data/ inside the zip file).
Experiment results for the RAG: results_{embedding_model}_{top-k}.json.
Experiment results for the Discovery Agent: results_{embedding_model}_{agent}_{refinement}_{llm}.json.
FAISS store (intermediate data required for exact reproduction of results; one folder for each embedding model): bge_small, nvidia and oai.
Intermediate data of the LLM-based refinement methods required for the exact reproduction of results: *_parser.json.
</dl
Data for: Olefin Metathesis under Spatial Confinement and Continuous Flow: Investigation of Isomeric Side Reactions with a Grubbs-Hoveyda Type Catalyst
No full textOlefin Metathesis under Spatial Confinement and Continuous Flow: We study olefin metathesis with a focus on isomers formed as by-products. For that purpose, a 2D-LC/MS system is implemented, showing decisive advantages over common 1H NMR analysis. As a consequence, it became possible to determine the influence of catalyst deactivation on both yield and selectivity of the reaction
Code for From Shapes to Shapes
No full textThis dataset contains the implementation code for an algorithm to infer SHACL shapes that the graph returned by an SPARQL CONSTRUCT query must satisfy if the input satisfies a given set of SHACL shapes. This dataset also includes an evaluation for the algorithm.
The algorithm implemented in this dataset is proposed in the paper From Shapes to Shapes: Inferring SHACL Shapes for Results of SPARQL CONSTRUCT Queries.
To execute the code, follow the instructions in the README.md file. For more info, please check the paper, and please have no hesitation to contact the authors for any inquiries.</p
D1244 sensor data (March 2024)
No full text
General information:
This dataset contains measurements from the adaptive high-rise demonstrator building D1244, built in the scope of the CRC1244. This 36m high building is equipped with 24 hydraulic actuators providing the basis for its structural adaptation. Strain gauges, pressure sensors and position encoders are mounted throughout the building and used for state estimation and monitoring.
Structure of the dataset:
Each zip-file contains measurements of one day in the hdf5 format.
The hdf5-files in each zip-file contain an array of 244 signals sampled over 10^6 time steps at approximately 200Hz.
labels.csv contains auxiliary information on all measured signals, including the sensor type and the sensor's location in the building
File contents:
Each hdf5-file contains signals of the following types, arranged as stated in labels.csv:
strain: strain (in mm/m) in columns and diagonal bracing elements, measured by strain gauges.
pressure: pressure (in bar) in the piston side or the rod side chamber of a hydraulic actuator. For actuators in the diagonal bracing, the rod side chamber is permanently connected to the tank.
posenc: displacement (in meters) of each actuator, measured by a position encoder.
optic: optically measured displacement (in meters) of emitters attached to the building's facade.
The building consists of four modules spanning three stories each, and all sensors within a module are connected to a control cabinet, from which all measurements are transmitted. The cameras of the optical measurement system are placed outside the building and transmit their data separately from the sensors within the building. Therefore, there are an additional two signals per module (or camera):
timestamp: unix timestamp (seconds since 1st January 1970) of the control cabinet
numvars: number of measured variables
Missing measurements are marked as NaN. The optical measurement system is currently undergoing maintenance, which is why the corresponding signals are all NaN.
</p
InteRead
No full textThe InteRead dataset is designed to explore the impact of interruptions on reading behavior. It includes eye-tracking data from 50 adults with normal or corrected-to-normal eyesight and proficiency in English (native or C1 level). The dataset encompasses a self-paced reading task of an English fictional text, with participants encountering interruptions and responding to unrelated questions. It features gaze data, cognitive scores, and demographic information. The dataset is a resource for interdisciplinary research in educational science, psycholinguistics, and natural language processing. More details can be found in the README.md
COLife_04 - Multi-Scale More-than-Human Atlas Gigamaping, Ecosystem Mapping and CoDesigning
No full textThe dataset presents the Multi-Scale More-than-Human Atlas Gigamapping Workshop output that should evaluate and suggest new developments of COLife project prototyping. We also gathered the ecological check data and planted new species with gardening and seedbombs. This was all elaborated in the workshop and speculated on within the 'Enacting Bateson's Ecological Aesthetics in Architecture and Design' team.
The workshop aimed to explore - from a multi-scale and transdisciplinary more-than-human perspective - the place-specific potential of the POL-AI, BioDiveln and NetWall prototypes in leveraging the Stuttgart University campus’s urban ecosystem. The investigation was conducted through the Multi-Scale Atlas (Porqueddu 2018, 2022, 2023) a tool for systemic enquiry which enables designers to (1) explore the cross-scale relationships in which place-specific spatial configurations are immersed, (2) frame them as parts of complex socio-material-environmental systems which evolve in a non-linear manner over time, (3) identify leverage points for design actions. The Atlas combines a multi-layer observation with multi-scale mapping with more-than-human perspective. This makes it possible to explore the relationships between seemingly separated perspectives and to intersect data collected through investigations from different disciplines, thus enhancing their potential synergies in fostering place-specific cycles of urban self-regeneration.
Further information can be found on the workshop website.</p
ABxM.MultiStorey.Columns: Agent-based Column Arrangement for Multi-Storey Structures
No full textABxM.MultiStorey.Columns is an add-on for the agent-based design and arrangement of columns in multi-storey, point-supported structures. The add-on contains various agent system constructs and utilities for column arrangement and is intended to be used within Rhino/Grasshopper. This version contains tools for the autonomous self-organisation of point-wise structural supports. The package includes example files that demonstrate how a model of an agent-based floor can arrange its own columns and exclude them from certain areas of the slab.
The database contains the source code for the ABxM.MultiStorey.Columns add-on to the ABxM Framework. By opening the ICL.sln solution in the Visual Studio IDE, both the ICL.Core and ICL.GH projects can be compiled to generate a grasshopper plugin for the add-in. The add-in contains three specialsied behaviours and a custom agent system, which are to be used in conjunction with the canonical ABxM Cartesian Agents and Cartesian Environment objects. Example file, in the form of a Rhino 3D BasePlan and a grasshopper script implementing the add-in are provided in the “data” folder
Results for pseudo-3D Stokes simulations with a geometry-informed drag term formulation for porous media with varying apertures
No full textContent:
This data set includes all the necessary input data to fully replicate the benchmarks and applications in Krach et al.(2024). In addition, the results for all solvers are provided as raw simulation output as well as in condense form.
Complete datasets: 01_CYLINDER, 02_SINGLE_PRECIPITATE, 03_SEGMENT1, 04_SEGMENT2, 05_SEGMENT3
Each data set consists of all relevant input data for poremaps and DuMux as well as the results of both solvers. A README contains relevant information such as voxelsize, geometry size and boundary conditions. Pseudo-3D input data and results comprise different folders, each containing a different drag formulation.
Input data for DuMux may be generated from the 3D input data using the localdrag python module. See the README or the source code for detailed information. For further details on the input files, please refer to the software documentation of the corresponding DuMux module Pseudo-3D-Stokes Module and poremaps. For input, evaluation or modification of the simulation results, I/O routines, interpolation algorithms and analysis tools are provided in the localdrag python module.
Condensed results:
06_EVALUATION_PERMEABILITIES
All permeabilities and relative errors for all domains can be found in condensed form (.csv files) in 06_EVALUATION_PERMEABILITIES. In order to plot this with the scripts provided, the functionalities provided in the localdrag python module are required.
Related datasets and repositories:
POREMAPS: git repository, DaRUS dataset
DuMux:
Website, git repository
Image dataset of micromodel with precipitation: DaRUS dataset, research paper
localdrag python module for preprocessing and data handling: git repository, DaRUS dataset
</ul
Floor Type Detection Dataset
No full textDataset for Floor Type Detection of the Robot Dog Unitree Go1 Edu. Details can be found in the seperate report.
!!Privacy Statement!!
This dataset is made available for academic use only. However, we take your privacy seriously! If you find yourself or personal belongings in this dataset and feel unwell about it, please contact us and we will immediately remove the respective data from our server
Replication Data for: Development of a framework for point-particle tracking
No full textThis dataset contains a custom build for Altair's simulation software ultraFluidX and the input files used in the Master Thesis 'Development of an Euler-Lagrangian framework for point-particle tracking to enable efficient multiscale simulations of complex flows', submitted by Helena Kschidock on 02/11/2023.
For further details on the input files, as well as user instructions for the custom build, please refer to the README.</p