Machine Learning and Knowledge Extraction in Digital Pathology Needs an Integrative Approach

AbstractThe precise targeting of antibodies and other protein therapeutics is required for their proper function and the elimination of deleterious off-target effects. Often the molecular structure of a therapeutic target is unknown and randomized methods are used to design antibodies without a model that relates antibody sequence to desired properties. Here we present a machine learning method that can design human Immunoglobulin G (IgG) antibodies with target affinities that are superior to candidates from phage display panning experiments within a limited design budget. We also demonstrate that machine learning can improve target-specificity by the modular composition of models from different experimental campaigns, enabling a new integrative approach to improving target specificity. Our results suggest a new path for the discovery of therapeutic molecules by demonstrating that predictive and differentiable models of antibody binding can be learned from high-throughput experimental data without the need for target structural data.SignificanceAntibody based therapeutics must meet both affinity and specificity metrics, and existing in vitro methods for meeting these metrics are based upon randomization and empirical testing. We demonstrate that with sufficient target-specific training data machine learning can suggest novel antibody variable domain sequences that are superior to those observed during training. Our machine learning method does not require any target structural information. We further show that data from disparate antibody campaigns can be combined by machine learning to improve antibody specificity.

Download Full-text

Using a DEA–AutoML Approach to Track SDG Achievements

Sustainability ◽

10.3390/su122310124 ◽

2020 ◽

Vol 12 (23) ◽

pp. 10124

Author(s):

Bodin Singpai ◽

Desheng Wu

Keyword(s):

Machine Learning ◽

Back Propagation ◽

Back Propagation Neural Network ◽

Correlated Data ◽

Integrative Approach ◽

Classification Prediction ◽

Predicted Values ◽

Highly Correlated ◽

Prediction Problems ◽

Data Result

Each country needs to monitor progress on their Sustainable Development Goals (SDGs) to develop strategies that meet the expectations of the United Nations. Data envelope analysis (DEA) can help identify best practices for SDGs by setting goals to compete against. Automated machine learning (AutoML) simplifies machine learning for researchers who need less time and manpower to predict future situations. This work introduces an integrative method that integrates DEA and AutoML to assess and predict performance in SDGs. There are two experiments with different data properties in their interval and correlation to demonstrate the approach. Three prediction targets are set to measure performance in the regression, classification, and multi-target regression algorithms. The back-propagation neural network (BPNN) is used to validate the outputs of the AutoML. As a result, AutoML can outperform BPNN for regression and classification prediction problems. Low standard deviation (SD) data result in poor prediction performance for the BPNN, but does not have a significant impact on AutoML. Highly correlated data result in a higher accuracy, but does not significantly affect the R-squared values between the actual and predicted values. This integrative approach can accurately predict the projected outputs, which can be used as national goals to transform an inefficient country into an efficient country.

Download Full-text

Towards Integrative Machine Learning and Knowledge Extraction

10.1007/978-3-319-69775-8 ◽

2017 ◽

Cited By ~ 1

Keyword(s):

Machine Learning ◽

Knowledge Extraction

Download Full-text

Towards an Integrative Approach for Automated Literature Reviews Using Machine Learning

Proceedings of the 53rd Hawaii International Conference on System Sciences ◽

10.24251/hicss.2020.095 ◽

2020 ◽

Author(s):

Christoph Tauchert ◽

Marco Bender ◽

Neda Mesbah ◽

Peter Buxmann

Keyword(s):

Machine Learning ◽

Integrative Approach ◽

Literature Reviews

Download Full-text

Information modeling and knowledge extraction for machine learning applications in industrial production systems

Machine Learning for Cyber Physical Systems - Technologien für die intelligente Automation ◽

10.1007/978-3-662-62746-4_8 ◽

2020 ◽

pp. 73-81

Author(s):

Stefan Windmann ◽

Christian Kühnert

Keyword(s):

Machine Learning ◽

Domain Knowledge ◽

Structural Information ◽

Semantic Annotation ◽

Information Model ◽

Knowledge Extraction ◽

Information Modeling ◽

Sensor Data ◽

Tool Chain ◽

Machine Learning Applications

AbstractIn this paper, a new information model for machine learning applications is introduced, which allows for a consistent acquisition and semantic annotation of process data, structural information and domain knowledge from industrial productions systems. The proposed information model is based on Industry 4.0 components and IEC 61360 component descriptions. To model sensor data, components of the OGC SensorThings model such as data streams and observations have been incorporated in this approach. Machine learning models can be integrated into the information model in terms of existing model serving frameworks like PMML or Tensorflowgraph. Based on the proposed information model, a tool chain for automatic knowledge extraction is introduced and the automatic classification of unstructured text is investigated as a particular application case for the proposed tool chain.

Download Full-text