A5024125738- Adversarial Robustness in Machine Learning
- Machine Learning and Data Classification
- Anomaly Detection Techniques and Applications
- Machine Learning and Algorithms
- Stochastic Gradient Optimization Techniques
- Rough Sets and Fuzzy Logic
- Advanced Neural Network Applications
- Neural Networks and Applications
- Data Mining Algorithms and Applications
- Statistical Methods and Inference
- Information Retrieval and Search Behavior
- Public Administration and Political Analysis
- Explainable Artificial Intelligence (XAI)
- Sociology and Education Studies
- Human Pose and Action Recognition
- Recommender Systems and Techniques
- Advanced Graph Neural Networks
- Product Development and Customization
- Topic Modeling
- Bayesian Modeling and Causal Inference
- Graph Theory and Algorithms
- Network Security and Intrusion Detection
- Web Data Mining and Analysis
- Big Data and Digital Economy
- Education, Psychology, and Social Research
University of Passau
2024
Heinz Nixdorf Stiftung
2024
Google (Switzerland)
2023
Google (United States)
2015-2020
University of Freiburg
2018
Learning-to-Rank deals with maximizing the utility of a list examples presented to user, items higher relevance being prioritized. It has several practical applications such as large-scale search, recommender systems, document summarization and question answering. While there is widespread support for classification regression based learning, learning-to-rank in deep learning been limited. We introduce TensorFlow Ranking, first open source library solving ranking problems framework. highly...
Real-world machine learning applications may have requirements beyond accuracy, such as fast evaluation times and interpretability. In particular, guaranteed monotonicity of the learned function with respect to some inputs can be critical for user confidence. We propose meeting these goals low-dimensional problems by flexible, monotonic functions using calibrated interpolated look-up tables. extend structural risk minimization framework lattice regression adding linear inequality...
Real-world machine learning applications may require functions that are fast-to-evaluate and interpretable. In particular, guaranteed monotonicity of the learned function can be critical to user trust. We propose meeting these goals for low-dimensional problems by flexible, monotonic using calibrated interpolated look-up tables. extend structural risk minimization framework lattice regression train tables solving a convex problem with appropriate linear inequality constraints. addition, we...
We propose learning deep models that are monotonic with respect to a user-specified set of inputs by alternating layers linear embeddings, ensembles lattices, and calibrators (piecewise functions), appropriate constraints for monotonicity, jointly training the resulting network. implement projections new computational graph nodes in TensorFlow use ADAM optimizer batched stochastic gradients. Experiments on benchmark real-world datasets show six-layer lattice networks achieve state-of-the art...
TensorFlow-GNN (TF-GNN) is a scalable library for Graph Neural Networks in TensorFlow. It designed from the bottom up to support kinds of rich heterogeneous graph data that occurs today's information ecosystems. In addition enabling machine learning researchers and advanced developers, TF-GNN offers low-code solutions empower broader developer community learning. Many production models at Google use TF-GNN, it has been recently released as an open source project. this paper we describe...
Yggdrasil Decision Forests is a library for the training, serving and interpretation of decision forest models, targeted both at research production work, implemented in C++, available command line interface, Python (under name TensorFlow Forests), JavaScript, Go, Google Sheets Simple ML Sheets). The has been developed organically since 2018 following set four design principles applicable to machine learning libraries frameworks: simplicity use, safety modularity high-level abstraction,...
Abstract Engineering projects for complex technical systems such as automobiles demand extensive requirement specifications and corresponding hierarchy levels in system architectures. Especially when considering emergent phenomena, total weight or aerodynamics, a closely networked collaboration of discipline‐specific cross‐disciplinary roles is required. Further, large organizations with group structure, resulting functional non‐functional contents need to be managed by distinct Systems...
Minimizing empirical risk subject to a set of constraints can be useful strategy for learning restricted classes functions, such as monotonic submodular classifiers that guarantee certain class label some subset examples, etc. However, these restrictions may result in very large number constraints. Projected stochastic gradient descent (SGD) is often the default choice large-scale optimization machine learning, but requires projection after each update. For heavily-constrained objectives, we...
Decision forest algorithms typically model data by learning a binary tree structure recursively where every node splits the feature space into two sub-regions, sending examples left or right branch as result. In axis-aligned decision forests, "decision" to route an input example is result of evaluation condition on single dimension in space. Such conditions are learned using efficient, often greedy that optimize local loss function. For example, node's may be threshold function applied...
Axis-aligned decision forests have long been the leading class of machine learning algorithms for modeling tabular data. In many applications such as learning-to-rank, deliver remarkable performance. They also possess other coveted characteristics interpretability. Despite their widespread use and rich history, to date fail consume raw structured data text, or learn effective representations them, a factor behind success deep neural networks in recent years. While there exist methods that...
Worldwide trainers ask if there is a rotation scheme, which improves the gymnastics performance and/or facilitates learning of elements with longitudinal rotations. Although are some surveys and scientific publications on it, we still seeking for more data to understand undergoing relationships within habits high-level gymnasts. In recent study, Men’s Individual All‐Around finalists at Olympic Games Rio 2016 were categorized using current classification system rotational schemes. This study...
Yggdrasil Decision Forests is a library for the training, serving and interpretation of decision forest models, targeted both at research production work, implemented in C++, available command line interface, Python (under name TensorFlow Forests), JavaScript, Go, Google Sheets Simple ML Sheets). The has been developed organically since 2018 following set four design principles applicable to machine learning libraries frameworks: simplicity use, safety modularity high-level abstraction,...