Projekt: Data Quality Research

Research in the field of "Data Science"

The Data Science +X Research Center at the BHT conducts use-inspired basic research.

Our use-inspired basic research is often based on public funding with industrial or health care partners. This basic research has a horizon of about 36-60 months and looks at fundamental and often risky research questions. Few projects also investigate to focus on very narrow and rather applied research questions. The typical project duration for these projects is about 12-18 months.

Appl-FM (DFG)

Berlin Initiative for Applied Foundation Model Research

Teilprojekt: Prädiktive Medizin: Zweitmeinung durch KI

Manipulating Virtual Self-Perception through Visuo-haptic Avatar Parameters (DFG)

Research into the combined effect of avatars’ visual presentations and haptic stimuli on users’ self-perception ...

Real-time Intersensory Discomfort Compensation (DFG)

Systematic research of multimodal sensory illusions in mixed reality.

Monitoring: Peace & Security (DSF)

Citizen-based Monitoring for Peace & Security in Era of Synthetic Media and Deepfakes

RIWWER (BMWi)

Reduction of the impact of untreated wastewater on the environment in case of torrential rain

KInsekt (BMUV)

KI basiertes Insektenmonitoring mit Citizen Science

VITALab.Mobile (BMBF)

A mobile VR/AR laboratory for case and field studies as well as clinical trials to evaluate novel forms of virtual therapies and scientific research.

Green Consumption Assistant (BMUV)

Artificial Intelligence (AI) based Assistant that will support consumers to make more sustainable shopping decisions

Berlin ZOO

Personalized online-tracking of spatial memory consolidation in a virtual environment

Human.VR.Lab (BMBF)

with focus on the interfaces between the life sciences, robotics, mechanical engineering and computer science

SynthNet (BMBF)

A solution approach consisting of synthetically generating all required comparison images from a real-world industry grade 3D CAD data and using them for object identification.

BewARe (BMBF)

Based on sensor data acquired from the patients during training sessions the system adapts to patients individual training needs and thereby serves as a non-drug therapy for people with hypertension.

Energy Cost Savings (EU-EFRE: BENE)

Forecasting Energy Cost Savings in Weather-Guided Heating Control

VR-REACH (IFAF)

A novel method for the therapy of school anxiety in children and adolescents using a VR application as part of a pilot study.

Transparency in Machine Learning

A collaboration with Philipp Schmidt, Amazon Research, and Prof. Timm Teubner, TU Berlin, to investigate whether and when transparency in AI actually increases trust in AI systems.

Maschinelles Lernen für Biodiversität (BMUV)

Entwicklung von Verfahren des Maschinellen Lernens (ML) zur automatisierten Erkennung von Insekten

Interaction-based RDW

RDW utilizes imperfections of human perception to introduce small changes like rotations or translations to steer the user away from the tracking boundaries.

Vessel Segmentation

Our work highly encourages the development of clinically applicable segmentation tools based on deep learning.

COMFORT

(EU-HORIZON-HLTH-2020 )

Computational models for patient stratification in urologic cancers – Creating robust and trustworthy multimodal AI for health care

Pervasive Touch (DFG)

Methods for designing robust microgestures for touch-based interaction

AnorexiaVR (BMG)

Virtual body exposure therapy for adolescent anorexia nervosa patients

AI4SCM (BMBF)

Machine Reading im Supply Chain Management für KMUs

HapticIO (BMBF)

With the help of the generic controller, haptic feedback can be generated for users even over distances...

More-with-Less (BMBF)

Efficient language models for SMEs

Servicemeister (BMWi)

Artificial intelligence (AI) based cross-company service platform for the german industry

PLASS (BMWi)

Platform for AI-based decision support for supply chain management

DiSEA (BMBF)

Digitale Studiengänge - Analyse von Erfolgs- und Abbruchfaktoren

Illusionary Surface Interfaces (DFG)

We aim to create novel interactive experiences, that exploit multisensory illusions in order to extend the range of interface properties that can be displayed, using only everyday object surfaces as interfaces.

WINK _ Gestensteuerung (BMBF)

Die Hände und Finger als primäre Werkzeuge in der physischen Welt sollen für die Interaktion mit digitalen Geräten nutzbar werden.

NOHATE (BMBF)

Examining hate communication in social media, online forums and commentary areas

Students Advice

to exploit the data that universities have about the academic achievements of their current and past students

FLFDDR (DFG)

Research platform "Literary field GDR": authors, works, networks – Pilot study

mEDUator (BMBF)

Mediatorplattform für interoperable Bildungstechnologien (mEDUator)

ExCell (BMWi)

Enableing integrated mobility services for medium-sized enterprises.

KIP-SDM (BMBF)

Datenintegrations- und Datenanalysemethoden sowie darauf aufbauende KI-Anwendungen für Sturzprävention – Beispielanwendung für alternative relevante Fragestellungen

PAN-Assistant (BMBF)

AR/VR supported orientation and memory training for Post-Acute Neurorehabilitation

Outcome Prediction from Patient Letters

AI-based outcome prediction from clinical texts

Tbox

A prediction model for permanent graft loss in renal transplant patients

Clinical Assertion Classification

Evaluation of state-of-the-art medical language models

Cross-Lingual Knowledge Transfer for Clinical Phenotyping

Automatic extraction of clinical conditions from patient records

A Quantitative Force Map of the Mitotic Spindle (DFG)

Understanding the way in which cells engineer micrometer-scale structures.

CELLbioprints (IFAF)

Mittels 3D-Druck zelluläre Strukturen zu simulieren und Beziehungen zwischen Materialeigenschaften und Geometrie etablieren

MACSS (BMWi)

Developing the prototype of a novel patient-centered Smart Health Service platform

Smart-MD (BMWi)

Converting medical case data from clinics into legally secure data products.

Data Quality in Machine Learning Systems

developing methods for better automation of monitoring of data quality, improvement of data quality and prediction of data quality problems in ML production systems