Anton Thielmann’s research while affiliated with Technische Universität Clausthal and other places

The analysis of tabular data has traditionally been dominated by gradient-boosted decision trees (GBDTs), known for their proficiency with mixed categorical and numerical features. However, recent deep learning innovations are challenging this dominance. We introduce Mambular, an adaptation of the Mamba architecture optimized for tabular data. We extensively benchmark Mambular against state-of-the-art models, including neural networks and tree-based methods, and demonstrate its competitive performance across diverse datasets. Additionally, we explore various adaptations of Mambular to understand its effectiveness for tabular data. We investigate different pooling strategies, feature interaction mechanisms, and bi-directional processing. Our analysis shows that interpreting features as a sequence and passing them through Mamba layers results in surprisingly performant models. The results highlight Mambulars potential as a versatile and powerful architecture for tabular data analysis, expanding the scope of deep learning applications in this domain. The source code is available at https://github.com/basf/mamba-tabular.

We present an Natural Language Processing based analysis on the phenomenon of “Meme Stocks”, which has emerged as a result of the proliferation of neo-brokers like Robinhood and the massive increase in the number of small-scale stock investors. Such investors often use specific Social Media channels to share short-term investment decisions and strategies, resulting in partial collusion and planning of investment decisions. The impact of online communities on the stock prices of affected companies has been considerable in the short term. This paper has two objectives. Firstly, we chronologically model the discourse on the most prominent platforms. Secondly, we examine the potential for using collaboratively made investment decisions as a means to assist in the selection of potential investments.. To understand the investment decision-making processes of small-scale investors, we analyze data from Social Media platforms like Reddit, Stocktwits and Seeking Alpha. Our methodology combines Sentiment Analysis and Topic Modelling. Sentiment Analysis is conducted using VADER and a fine-tuned BERT model. For Topic Modelling, we utilize LDA, NMF and the state-of-the-art BERTopic. We identify the topics and shapes of discussions over time and evaluate the potential for leveraging information of the decision-making process of investors for trading choices. We utilize Random Forest and Neural Network Models to show that latent information in discussions can be exploited for trend prediction of stocks affected by Social Network driven herd behavior. Our findings provide valuable insights into content and sentiment of discussions and are a vehicle to improve efficient trading decisions for stocks affected from short-term herd behavior.

Extracting and identifying latent topics in large text corpora has gained increasing importance in Natural Language Processing (NLP). Most models, whether probabilistic models similar to Latent Dirichlet Allocation (LDA) or neural topic models, follow the same underlying approach of topic interpretability and topic extraction. We propose a method that incorporates a deeper understanding of both sentence and document themes, and goes beyond simply analyzing word frequencies in the data. Through simple corpus expansion, our model can detect latent topics that may include uncommon words or neologisms, as well as words not present in the documents themselves. Additionally, we propose several new evaluation metrics based on intruder words and similarity measures in the semantic space. We present correlation coefficients with human identification of intruder words and achieve near-human level results at the word-intrusion task. We demonstrate the competitive performance of our method with a large benchmark study, and achieve superior results compared to state-of-the-art topic modeling and document clustering models.

Extracting and identifying latent topics in large text corpora has gained increasing importance in Natural Language Processing (NLP). Most models, whether probabilistic models similar to Latent Dirichlet Allocation (LDA) or neural topic models, follow the same underlying approach of topic interpretability and topic extraction. We propose a method that incorporates a deeper understanding of both sentence and document themes, and goes beyond simply analyzing word frequencies in the data. This allows our model to detect latent topics that may include uncommon words or neologisms, as well as words not present in the documents themselves. Additionally, we propose several new evaluation metrics based on intruder words and similarity measures in the semantic space. We present correlation coefficients with human identification of intruder words and achieve near-human level results at the word-intrusion task. We demonstrate the competitive performance of our method with a large benchmark study, and achieve superior results compared to state-of-the-art topic modeling and document clustering models.

Deep neural networks (DNNs) have shown exceptional performances in a wide range of tasks and have become the go-to method for problems requiring high-level predictive power. There has been extensive research on how DNNs arrive at their decisions, however, the inherently uninterpretable networks remain up to this day mostly unobservable "black boxes". In recent years, the field has seen a push towards interpretable neural networks, such as the visually interpretable Neural Additive Models (NAMs). We propose a further step into the direction of intelligibility beyond the mere visualization of feature effects and propose Structural Neural Additive Models (SNAMs). A modeling framework that combines classical and clearly interpretable statistical methods with the predictive power of neural applications. Our experiments validate the predictive performances of SNAMs. The proposed framework performs comparable to state-of-the-art fully connected DNNs and we show that SNAMs can even outperform NAMs while remaining inherently more interpretable.

Deep neural networks (DNNs) have proven to be highly effective in a variety of tasks, making them the go-to method for problems requiring high-level predictive power. Despite this success, the inner workings of DNNs are often not transparent, making them difficult to interpret or understand. This lack of interpretability has led to increased research on inherently interpretable neural networks in recent years. Models such as Neural Additive Models (NAMs) achieve visual interpretability through the combination of classical statistical methods with DNNs. However, these approaches only concentrate on mean response predictions, leaving out other properties of the response distribution of the underlying data. We propose Neural Additive Models for Location Scale and Shape (NAMLSS), a modelling framework that combines the predictive power of classical deep learning models with the inherent advantages of distributional regression while maintaining the interpretability of additive models.

Mixture Density Networks (MDN) belong to a class of models that can be applied to data which cannot be sufficiently described by a single distribution since it originates from different components of the main unit and therefore needs to be described by a mixture of densities. In some situations, however, MDNs seem to have problems with the proper identification of the latent components. While these identification issues can to some extent be contained by using custom initialization strategies for the network weights, this solution is still less than ideal since it involves subjective opinions. We therefore suggest replacing the hidden layers between the model input and the output parameter vector of MDNs and estimating the respective distributional parameters with penalized cubic regression splines. Applying this approach to data from Gaussian mixture distributions as well gamma mixture distributions proved to be successful with the identification issues not playing a role anymore and the splines reliably converging to the true parameter values.

Topic models are a useful and popular method to find latent topics of documents. However, the short and sparse texts in social media micro-blogs such as Twitter are challenging for the most commonly used Latent Dirichlet Allocation (LDA) topic model. We compare the performance of the standard LDA topic model with the Gibbs Sampler Dirichlet Multinomial Model (GSDMM) and the Gamma Poisson Mixture Model (GPM), which are specifically designed for sparse data. To compare the performance of the three models, we propose the simulation of pseudo-documents as a novel evaluation method. In a case study with short and sparse text, the models are evaluated on tweets filtered by keywords relating to the Covid-19 pandemic. We find that standard coherence scores that are often used for the evaluation of topic models perform poorly as an evaluation metric. The results of our simulation-based approach suggest that the GSDMM and GPM topic models may generate better topics than the standard LDA model.