On Representation Learning for Scientific News Articles Using Heterogeneous Knowledge Graphs

In the era of misinformation and information inflation, the credibility assessment of the produced news is of the essence. However, fact-checking can be challenging considering the limited references presented in the news. This challenge can be transcended by utilizing the knowledge graph that is related to the news articles. In this work, we present a methodology for creating scientific news article representations by modeling the directed graph between the scientific news articles and the cited scientific publications. The network used for the experiments is comprised of the scientific news articles, their topic, the cited research literature, and their corresponding authors. We implement and present three different approaches: 1) a baseline Relational Graph Convolutional Network (R-GCN), 2) a Heterogeneous Graph Neural Network (HetGNN) and 3) a Heterogeneous Graph Transformer (HGT). We test these models in the downstream task of link prediction on the: a) news article - paper links and b) news article - article topic links. The results show promising applications of graph neural network approaches in the domains of knowledge tracing and scientific news credibility assessment.

Genetic Representation of Adaptive Neural Controllers

Peter Dürr

The manual design of adaptive controllers for robotic systems that face unpredictable environmental changes is often challenging. There is thus a growing interest in the development of automatic design tools to assist control engineers. One of the most common approaches in this domain is the evolutionary synthesis of Artificial Neural Networks (ANNs), which allows the automatic design of control systems with little or no human intervention. The performance of evolved neural controllers, however, critically depends on the choice of a suitable genetic representation, i.e., a description of the ANN on which the evolutionary search can operate effectively. In this thesis, I propose an alternative to existing representations for neural controllers based on an extension of Analog Genetic Encoding (AGE), an abstraction of the regulatory networks that control the expression of genes in biological organisms. The proposed approach unites three characteristics that have separately been considered advantageous, but have so far not been aggregated into a single representation: (1) it can be used to effectively synthesize the topology, weights, and numerical parameters of arbitrary networks; (2) it permits the use of heterogeneous ANN models that feature multiple types of signals; and (3) it facilitates the evolution of modular controllers. The results of a series of experiments demonstrate the advantages of these characteristics: (i) neural architectures synthesized with the proposed method display a performance competitive to the best hand-designed networks and networks synthesized with other representations; (ii) in tasks that feature an unpredictable changing environment, controllers with a heterogeneous ANN model outperform controllers with conventional, homogeneous ANN models; and (iii) in tasks that feature multiple modular sub-problems, the proposed representation allows the automatic decomposition of the problem, leading to a substantial improvement in the performance of the evolved controllers. In summary, the proposed representation in combination with a heterogeneous ANN model and an evolutionary algorithm represents a promising method for the automatic synthesis of adaptive control systems for robots.

EPFL2011

Biologically plausible unsupervised learning in shallow and deep neural networks

Bernd Albert Illing

The way our brain learns to disentangle complex signals into unambiguous concepts is fascinating but remains largely unknown. There is evidence, however, that hierarchical neural representations play a key role in the cortex. This thesis investigates biologically plausible models of unsupervised learning of hierarchical representations as found in the brain and modern computer vision models. We use computational modeling to address three main questions at the intersection of artificial intelligence (AI) and computational neuroscience.The first question is: What are useful neural representations and when are deep hierarchical representations needed? We approach this point with a systematic study of biologically plausible unsupervised feature learning in a shallow 2-layer networks on digit (MNIST) and object (CIFAR10) classification. Surprisingly, random features support high performance, especially for large hidden layers. When combined with localized receptive fields, random feature networks approach the performance of supervised backpropagation on MNIST, but not on CIFAR10. We suggest that future models of biologically plausible learning should outperform such random feature benchmarks on MNIST, or that such models should be evaluated in different ways.The second question is: How can hierarchical representations be learned with mechanisms supported by neuroscientific evidence? We cover this question by proposing a unifying Hebbian model, inspired by common models of V1 simple and complex cells based on unsupervised sparse coding and temporal invariance learning. In shallow 2-layer networks, our model reproduces learning of simple and complex cell receptive fields, as found in V1. In deeper networks, we stack multiple layers of Hebbian learning but find that it does not yield hierarchical representations of increasing usefulness. From this, we hypothesise that standard Hebbian rules are too constrained to build increasingly useful representations, as observed in higher areas of the visual cortex or deep artificial neural networks.The third question is: Can AI inspire learning models that build deep representations and are still biologically plausible? We address this question by proposing a learning rule that takes inspiration from neuroscience and recent advances in self-supervised deep learning. The proposed rule is Hebbian, i.e. only depends on pre- and post-synaptic neuronal activity, but includes additional local factors, namely predictive dendritic input and widely broadcasted modulation factors. Algorithmically, this rule applies self-supervised contrastive predictive learning to a causal, biological setting using saccades. We find that networks trained with this generalised Hebbian rule build deep hierarchical representations of images, speech and video.We see our modeling as a potential starting point for both, new hypotheses, that can be tested experimentally, and novel AI models that could benefit from added biological realism.

EPFL2021

Biologically plausible unsupervised learning in shallow and deep neural networks

Bernd Albert Illing

EPFL2021

Genetic Representation of Adaptive Neural Controllers

Peter Dürr

EPFL2011