Showing 1–7 of 7 results for author: Soelch, M

Integrating Competency-Based Education in Interactive Learning Systems

Authors: Maximilian Sölch, Moritz Aberle, Stephan Krusche

Abstract: Artemis is an interactive learning system that organizes courses, hosts lecture content and interactive exercises, conducts exams, and creates automatic assessments with individual feedback. Research shows that students have unique capabilities, previous experiences, and expectations. However, the course content on current learning systems, including Artemis, is not tailored to a student's compete… ▽ More Artemis is an interactive learning system that organizes courses, hosts lecture content and interactive exercises, conducts exams, and creates automatic assessments with individual feedback. Research shows that students have unique capabilities, previous experiences, and expectations. However, the course content on current learning systems, including Artemis, is not tailored to a student's competencies. The main goal of this paper is to describe how to make Artemis capable of competency-based education and provide individual course content based on the unique characteristics of every student. We show how instructors can define relations between competencies to create a competency relation graph, how Artemis measures and visualizes the student's progress toward mastering a competency, and how the progress can generate a personalized learning path for students that recommends relevant learning resources. Finally, we present the results of a user study regarding the usability of the newly designed competency visualization and give an outlook on possible improvements and future visions. △ Less

Submitted 25 August, 2023; originally announced September 2023.

Comments: 4 pages, 2 figures. Best Practitioner Report Award

Journal ref: Companion Proceedings of the 13th International Learning Analytics and Knowledge Conference (LAK'23), March 2023, pp. 53-56

Mind the Gap when Conditioning Amortised Inference in Sequential Latent-Variable Models

Authors: Justin Bayer, Maximilian Soelch, Atanas Mirchev, Baris Kayalibay, Patrick van der Smagt

Abstract: Amortised inference enables scalable learning of sequential latent-variable models (LVMs) with the evidence lower bound (ELBO). In this setting, variational posteriors are often only partially conditioned. While the true posteriors depend, e.g., on the entire sequence of observations, approximate posteriors are only informed by past observations. This mimics the Bayesian filter -- a mixture of smo… ▽ More Amortised inference enables scalable learning of sequential latent-variable models (LVMs) with the evidence lower bound (ELBO). In this setting, variational posteriors are often only partially conditioned. While the true posteriors depend, e.g., on the entire sequence of observations, approximate posteriors are only informed by past observations. This mimics the Bayesian filter -- a mixture of smoothing posteriors. Yet, we show that the ELBO objective forces partially-conditioned amortised posteriors to approximate products of smoothing posteriors instead. Consequently, the learned generative model is compromised. We demonstrate these theoretical findings in three scenarios: traffic flow, handwritten digits, and aerial vehicle dynamics. Using fully-conditioned approximate posteriors, performance improves in terms of generative modelling and multi-step prediction. △ Less

Submitted 17 March, 2021; v1 submitted 18 January, 2021; originally announced January 2021.

Comments: Published as a conference paper at ICLR 2021 (Poster)

Variational Tracking and Prediction with Generative Disentangled State-Space Models

Authors: Adnan Akhundov, Maximilian Soelch, Justin Bayer, Patrick van der Smagt

Abstract: We address tracking and prediction of multiple moving objects in visual data streams as inference and sampling in a disentangled latent state-space model. By encoding objects separately and including explicit position information in the latent state space, we perform tracking via amortized variational Bayesian inference of the respective latent positions. Inference is implemented in a modular neur… ▽ More We address tracking and prediction of multiple moving objects in visual data streams as inference and sampling in a disentangled latent state-space model. By encoding objects separately and including explicit position information in the latent state space, we perform tracking via amortized variational Bayesian inference of the respective latent positions. Inference is implemented in a modular neural framework tailored towards our disentangled latent space. Generative and inference model are jointly learned from observations only. Comparing to related prior work, we empirically show that our Markovian state-space assumption enables faithful and much improved long-term prediction well beyond the training horizon. Further, our inference model correctly decomposes frames into objects, even in the presence of occlusions. Tracking performance is increased significantly over prior art. △ Less

Submitted 14 October, 2019; originally announced October 2019.

On Deep Set Learning and the Choice of Aggregations

Authors: Maximilian Soelch, Adnan Akhundov, Patrick van der Smagt, Justin Bayer

Abstract: Recently, it has been shown that many functions on sets can be represented by sum decompositions. These decompositons easily lend themselves to neural approximations, extending the applicability of neural nets to set-valued inputs---Deep Set learning. This work investigates a core component of Deep Set architecture: aggregation functions. We suggest and examine alternatives to commonly used aggreg… ▽ More Recently, it has been shown that many functions on sets can be represented by sum decompositions. These decompositons easily lend themselves to neural approximations, extending the applicability of neural nets to set-valued inputs---Deep Set learning. This work investigates a core component of Deep Set architecture: aggregation functions. We suggest and examine alternatives to commonly used aggregation functions, including learnable recurrent aggregation functions. Empirically, we show that the Deep Set networks are highly sensitive to the choice of aggregation functions: beyond improved performance, we find that learnable aggregations lower hyper-parameter sensitivity and generalize better to out-of-distribution input size. △ Less

Submitted 8 April, 2020; v1 submitted 18 March, 2019; originally announced March 2019.

Approximate Bayesian inference in spatial environments

Authors: Atanas Mirchev, Baris Kayalibay, Maximilian Soelch, Patrick van der Smagt, Justin Bayer

Abstract: Model-based approaches bear great promise for decision making of agents interacting with the physical world. In the context of spatial environments, different types of problems such as localisation, map**, navigation or autonomous exploration are typically adressed with specialised methods, often relying on detailed knowledge of the system at hand. We express these tasks as probabilistic inferen… ▽ More Model-based approaches bear great promise for decision making of agents interacting with the physical world. In the context of spatial environments, different types of problems such as localisation, map**, navigation or autonomous exploration are typically adressed with specialised methods, often relying on detailed knowledge of the system at hand. We express these tasks as probabilistic inference and planning under the umbrella of deep sequential generative models. Using the frameworks of variational inference and neural networks, our method inherits favourable properties such as flexibility, scalability and the ability to learn from data. The method performs comparably to specialised state-of-the-art methodology in two distinct simulated environments. △ Less

Submitted 20 June, 2019; v1 submitted 18 May, 2018; originally announced May 2018.

Comments: Preprint of publication at RSS 2019

Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Data

Authors: Maximilian Karl, Maximilian Soelch, Justin Bayer, Patrick van der Smagt

Abstract: We introduce Deep Variational Bayes Filters (DVBF), a new method for unsupervised learning and identification of latent Markovian state space models. Leveraging recent advances in Stochastic Gradient Variational Bayes, DVBF can overcome intractable inference distributions via variational inference. Thus, it can handle highly nonlinear input data with temporal and spatial dependencies such as image… ▽ More We introduce Deep Variational Bayes Filters (DVBF), a new method for unsupervised learning and identification of latent Markovian state space models. Leveraging recent advances in Stochastic Gradient Variational Bayes, DVBF can overcome intractable inference distributions via variational inference. Thus, it can handle highly nonlinear input data with temporal and spatial dependencies such as image sequences without domain knowledge. Our experiments show that enabling backpropagation through transitions enforces state space assumptions and significantly improves information content of the latent embedding. This also enables realistic long-term prediction. △ Less

Submitted 3 March, 2017; v1 submitted 20 May, 2016; originally announced May 2016.

Comments: Published as a conference paper at ICLR 2017

Variational Inference for On-line Anomaly Detection in High-Dimensional Time Series

Authors: Maximilian Soelch, Justin Bayer, Marvin Ludersdorfer, Patrick van der Smagt

Abstract: Approximate variational inference has shown to be a powerful tool for modeling unknown complex probability distributions. Recent advances in the field allow us to learn probabilistic models of sequences that actively exploit spatial and temporal structure. We apply a Stochastic Recurrent Network (STORN) to learn robot time series data. Our evaluation demonstrates that we can robustly detect anomal… ▽ More Approximate variational inference has shown to be a powerful tool for modeling unknown complex probability distributions. Recent advances in the field allow us to learn probabilistic models of sequences that actively exploit spatial and temporal structure. We apply a Stochastic Recurrent Network (STORN) to learn robot time series data. Our evaluation demonstrates that we can robustly detect anomalies both off- and on-line. △ Less

Submitted 14 June, 2016; v1 submitted 23 February, 2016; originally announced February 2016.

Comments: Accepted as workshop paper at ICLR 2016; accepted as workshop paper for anomaly detection workshop at ICML 2016