Project
Small Innovative Projects
This project “small innovative project” (SIPs) organizes the selection, assessment, monitoring and reporting of SIP projects in the DDHT program. SIP projects are small projects in which new innovative ideas that do not fit into the current research are tested and are by definition short projects (< 1 year). There are 2 types of SIPs: i) Institute SIP (WR-SIP) and ii) specific/ dedicated SIPs in which cooperation between WR units is realized. The topics for the specific/ dedicated SIPs have been selected from a broad internal WUR consultation. For 2021, 5 WR-SIPS and 5 specific SIPs have been launched. The topics in the SIPs address 1 or more pillars in the DDHT research program.
The following SIPs are performed:
WR SIPs:
1. Federate learning in food safety; towards the development of a FAIR Data Train Infrastructure (WR unit: WFSR)
2. Automated break down analysis of time series for developments in Dutch agriculture (WR unit: WEcR)
3. Combining mixed-integer linear programming with machine learning for fast re-planning in the integrated-biomass-logistics-center based supply chains (WR unit: WFBR)
4. Improvement of pose estimation from poultry video data (WR unit: WLR).
5. Unlocking high-throughput phenotyping data from NPEC (WR Unit: WPR)
Specific SIPs
1. Advanced Machine Learning (WR units: WFSR, WBFR). While machine learning is already at the core of current DDHT projects, the following new methods have not yet received attention and have transformed the ML field in recent years: transformer networks, graphical neural networks, temporal convolutional networks, and knowledge graph embedding. The SIP will address this for some WUR domains.
2. Distributed computing (WR units: WENR, WPR, WLR). Calculations are usually no longer performed on a single node, but are performed in a network of cloud servers, mobile phones, smart sensors or using the unused computing resources of the network of computing equipment, etc. There are relevant developments in the more efficient use of the computing resources of a network of devices. The SIP will demonstrate such an approach in a recognized WUR domain.
3. Digital knowledge streams (WR units: WPR, WBFR, WLR). Traditionally, WUR knowledge is transferred through publications and reports. Applications or services that make models, data or expertise available can lead to new business models for WUR. The SIP will develop robust, scalable, product-level digital services that reveal our organization's needs in this area.
4. Natural language processing (WR units: WFSR, WEcR). While most of our data science activities involve numerical data (including coded images), text-based analysis is equally important. The SIP will demonstrate the use of advanced Natural Language Processing (NLP) techniques and tools in the agri-food domain.
5. Impact of digital technologies (WR units: WENR, WEcR)- Digital technologies often promise significant improvements or even disruption. However, the actual contribution of these technologies to societal goals (i.e. SDGs, economic developments, environmental performance and social engagement) remains unclear. The SIP will investigate this.