Frieder Stolzenburg

Stefanie Krause, Frieder Stolzenburg

Commonsense reasoning is a difficult task for a computer, but a critical skill for an artificial intelligence (AI). It can enhance the explainability of AI models by enabling them to provide intuitive and human-like explanations for their decisions. This is necessary in many areas especially in question answering (QA), which is one of the most important tasks of natural language processing (NLP). Over time, a multitude of methods have emerged for solving commonsense reasoning problems such as knowledge-based approaches using formal logic or linguistic analysis. In this paper, we investigate the effectiveness of large language models (LLMs) on different QA tasks with a focus on their abilities in reasoning and explainability. We study three LLMs: GPT-3.5, Gemma and Llama 3. We further evaluate the LLM results by means of a questionnaire. We demonstrate the ability of LLMs to reason with commonsense as the models outperform humans on different datasets. While GPT-3.5's accuracy ranges from 56% to 93% on various QA benchmarks, Llama 3 achieved a mean accuracy of 90% on all eleven datasets. Thereby Llama 3 is outperforming humans on all datasets with an average 21% higher accuracy over ten datasets. Furthermore, we can appraise that, in the sense of explainable artificial intelligence (XAI), GPT-3.5 provides good explanations for its decisions. Our questionnaire revealed that 66% of participants rated GPT-3.5's explanations as either "good" or "excellent". Taken together, these findings enrich our understanding of current LLMs and pave the way for future investigations of reasoning and explainability.

Cedric Haufe, Frieder Stolzenburg

We present a neurosymbolic approach, i.e., combining symbolic and subsymbolic artificial intelligence, to validating offer documents in regulated public institutions. We employ a language model to extract information and then aggregate with an LTN (Logic Tensor Network) to make an auditable decision. In regulated public institutions, decisions must be made in a manner that is both factually correct and legally verifiable. Our neurosymbolic approach allows existing domain-specific knowledge to be linked to the semantic text understanding of language models. The decisions resulting from our pipeline can be justified by predicate values, rule truth values, and corresponding text passages, which enables rule checking based on a real corpus of offer documents. Our experiments on a real corpus show that the proposed pipeline achieves performance comparable to existing models, while its key advantage lies in its interpretability, modular predicate extraction, and explicit support for XAI (Explainable AI).

Stefanie Krause, Oliver Otto, Frieder Stolzenburg

Recurrent neural networks are a powerful means in diverse applications. We show that, together with so-called conceptors, they also allow fast learning, in contrast to other deep learning methods. In addition, a relatively small number of examples suffices to train neural networks with high accuracy. We demonstrate this with two applications, namely speech recognition and detecting car driving maneuvers. We improve the state of the art by application-specific preparation techniques: For speech recognition, we use mel frequency cepstral coefficients leading to a compact representation of the frequency spectra, and detecting car driving maneuvers can be done without the commonly used polynomial interpolation, as our evaluation suggests.

Claudia Schon, Sophie Siebert, Frieder Stolzenburg

Negation is both an operation in formal logic and in natural language by which a proposition is replaced by one stating the opposite, as by the addition of "not" or another negation cue. Treating negation in an adequate way is required for cognitive reasoning, which aims at modeling the human ability to draw meaningful conclusions despite incomplete and inconsistent knowledge. One task of cognitive reasoning is answering questions given by sentences in natural language. There are tools based on discourse representation theory to convert sentences automatically into a formal logic representation, and additional knowledge can be added using the predicate names in the formula and knowledge databases. However, the knowledge in logic databases in practice always is incomplete. Hence, forward reasoning of automated reasoning systems alone does not suffice to derive answers to questions because, instead of complete proofs, often only partial positive knowledge can be derived, while negative knowledge is used only during the reasoning process. In consequence, we aim at eliminating syntactic negation, strictly speaking, the negated event or property. In this paper, we describe an effective procedure to determine the negated event or property in order to replace it by its inverse. This lays the basis of cognitive reasoning, employing both logic and machine learning for general question answering. We evaluate our procedure by several benchmarks and demonstrate its practical usefulness in our cognitive reasoning system.

Maria Heinze, Lars Hausfeld, Rainer Goebel et al.

An acoustic stimulus, e.g., a musical harmony, is transformed in a highly non-linear way during the hearing process in ear and brain. We study this by comparing the frequency spectrum of an input stimulus and its response spectrum in the auditory processing stream using the frequency following response (FFR). Using electroencephalography (EEG), we investigate whether the periodicity pitches of complex harmonies (which are related to their missing fundamentals) are added in the auditory brainstem by analyzing the FFR. While other experiments focus on common musical harmonies like the major and the minor triad and dyads, we also consider the suspended chord. The suspended chord causes tension foreign to the common triads and therefore holds a special role among the triads. While watching a muted nature documentary, the participants hear synthesized classic piano triads and single tones with a duration of 300ms for the stimulus and 100ms interstimulus interval. We acquired EEG data of 64 electrodes with a sampling rate of 5kHz to get a detailed enough resolution of the perception process in the human brain. Applying a fast Fourier transformation (FFT) on the EEG response, starting 50ms after stimulus onset, the evaluation of the frequency spectra shows that the periodicity pitch frequencies calculated beforehand +/-3Hz occur with some accuracy. However, jitter turned out as a problem here. Note that the sought-for periodicity pitch frequencies do not physically exist in the frequency spectra of the stimuli.

Claudia Schon, Sophie Siebert, Frieder Stolzenburg

The CoRg system is a system to solve commonsense reasoning problems. The core of the CoRg system is the automated theorem prover Hyper that is fed with large amounts of background knowledge. This background knowledge plays a crucial role in solving commonsense reasoning problems. In this paper we present different ways to use knowledge graphs as background knowledge and discuss challenges that arise.

Frieder Stolzenburg, Sandra Litz, Olivia Michael et al.

Recurrent neural networks are a powerful means to cope with time series. We show how autoregressive linear, i.e., linearly activated recurrent neural networks (LRNNs) can approximate any time-dependent function f(t). The approximation can effectively be learned by simply solving a linear equation system; no backpropagation or similar methods are needed. Furthermore, and this is the main contribution of this paper, the size of an LRNN can be reduced significantly in one step after inspecting the spectrum of the network transition matrix, i.e., its eigenvalues, by taking only the most relevant components. Therefore, in contrast to other approaches, we do not only learn network weights but also the network architecture. LRNNs have interesting properties: They end up in ellipse trajectories in the long run and allow the prediction of further values and compact representations of functions. We demonstrate this by several case studies, among them multiple superimposed oscillators (MSO), robotic soccer (RoboCup), and stock price prediction. LRNNs outperform the previous state-of-the-art for the MSO task with a minimal number of units.

Olivia Michael, Oliver Obst, Falk Schmidsberger et al.

RoboCup is an international scientific robot competition in which teams of multiple robots compete against each other. Its different leagues provide many sources of robotics data, that can be used for further analysis and application of machine learning. This paper describes a large dataset from games of some of the top teams (from 2016 and 2017) in RoboCup Soccer Simulation League (2D), where teams of 11 robots (agents) compete against each other. Overall, we used 10 different teams to play each other, resulting in 45 unique pairings. For each pairing, we ran 25 matches (of 10mins), leading to 1125 matches or more than 180 hours of game play. The generated CSV files are 17GB of data (zipped), or 229GB (unzipped). The dataset is unique in the sense that it contains both the ground truth data (global, complete, noise-free information of all objects on the field), as well as the noisy, local and incomplete percepts of each robot. These data are made available as CSV files, as well as in the original soccer simulator formats.

Olivia Michael, Oliver Obst, Falk Schmidsberger et al.

We present a new approach for identifying situations and behaviours, which we call "moves", from soccer games in the 2D simulation league. Being able to identify key situations and behaviours are useful capabilities for analysing soccer matches, anticipating opponent behaviours to aid selection of appropriate tactics, and also as a prerequisite for automatic learning of behaviours and policies. To support a wide set of strategies, our goal is to identify situations from data, in an unsupervised way without making use of pre-defined soccer specific concepts such as "pass" or "dribble". The recurrent neural networks we use in our approach act as a high-dimensional projection of the recent history of a situation on the field. Similar situations, i.e., with similar histories, are found by clustering of network states. The same networks are also used to learn so-called conceptors, that are lower-dimensional manifolds that describe trajectories through a high-dimensional state space that enable situation-specific predictions from the same neural network. With the proposed approach, we can segment games into sequences of situations that are learnt in an unsupervised way, and learn conceptors that are useful for the prediction of the near future of the respective situation.

Frieder Stolzenburg, Florian Ruh

The fields of neural computation and artificial neural networks have developed much in the last decades. Most of the works in these fields focus on implementing and/or learning discrete functions or behavior. However, technical, physical, and also cognitive processes evolve continuously in time. This cannot be described directly with standard architectures of artificial neural networks such as multi-layer feed-forward perceptrons. Therefore, in this paper, we will argue that neural networks modeling continuous time are needed explicitly for this purpose, because with them the synthesis and analysis of continuous and possibly periodic processes in time are possible (e.g. for robot behavior) besides computing discrete classification functions (e.g. for logical reasoning). We will relate possible neural network architectures with (hybrid) automata models that allow to express continuous processes.

Ulrich Furbach, Claudia Schon, Frieder Stolzenburg et al.

Higher-level cognition includes logical reasoning and the ability of question answering with common sense. The RatioLog project addresses the problem of rational reasoning in deep question answering by methods from automated deduction and cognitive computing. In a first phase, we combine techniques from information retrieval and machine learning to find appropriate answer candidates from the huge amount of text in the German version of the free encyclopedia "Wikipedia". In a second phase, an automated theorem prover tries to verify the answer candidates on the basis of their logical representations. In a third phase - because the knowledge may be incomplete and inconsistent -, we consider extensions of logical reasoning to improve the results. In this context, we work toward the application of techniques from human reasoning: We employ defeasible reasoning to compare the answers w.r.t. specificity, deontic logic, normative reasoning, and model construction. Moreover, we use integrated case-based reasoning and machine learning techniques on the basis of the semantic structure of the questions and answer candidates to learn giving the right answers.

Ulrich Furbach, Claudia Schon, Frieder Stolzenburg

Deontic logic is a very well researched branch of mathematical logic and philosophy. Various kinds of deontic logics are considered for different application domains like argumentation theory, legal reasoning, and acts in multi-agent systems. In this paper, we show how standard deontic logic can be used to model ethical codes for multi-agent systems. Furthermore we show how Hyper, a high performance theorem prover, can be used to prove properties of these ethical codes.

Ulrich Furbach, Claudia Schon, Frieder Stolzenburg

This paper briefly characterizes the field of cognitive computing. As an exemplification, the field of natural language question answering is introduced together with its specific challenges. A possibility to master these challenges is illustrated by a detailed presentation of the LogAnswer system, which is a successful representative of the field of natural language question answering.

Ulrich Furbach, Claudia Schon, Frieder Stolzenburg

Deontic logic is a very well researched branch of mathematical logic and philosophy. Various kinds of deontic logics are discussed for different application domains like argumentation theory, legal reasoning, and acts in multi-agent systems. In this paper, we show how standard deontic logic can be stepwise transformed into description logic and DL- clauses, such that it can be processed by Hyper, a high performance theorem prover which uses a hypertableau calculus. Two use cases, one from multi-agent research and one from the development of normative system are investigated.

Claus-Peter Wirth, Frieder Stolzenburg

In the middle of the 1980s, David Poole introduced a semantical, model-theoretic notion of specificity to the artificial-intelligence community. Since then it has found further applications in non-monotonic reasoning, in particular in defeasible reasoning. Poole tried to approximate the intuitive human concept of specificity, which seems to be essential for reasoning in everyday life with its partial and inconsistent information. His notion, however, turns out to be intricate and problematic, which --- as we show --- can be overcome to some extent by a closer approximation of the intuitive human concept of specificity. Besides the intuitive advantages of our novel specificity ordering over Poole's specificity relation in the classical examples of the literature, we also report some hard mathematical facts: Contrary to what was claimed before, we show that Poole's relation is not transitive. The present means to decide our novel specificity relation, however, show only a slight improvement over the known ones for Poole's relation, and further work is needed in this aspect.

Frieder Stolzenburg

The perception of consonance/dissonance of musical harmonies is strongly correlated with their periodicity. This is shown in this article by consistently applying recent results from psychophysics and neuroacoustics, namely that the just noticeable difference between pitches for humans is about 1% for the musically important low frequency range and that periodicities of complex chords can be detected in the human brain. Based thereon, the concepts of relative and logarithmic periodicity with smoothing are introduced as powerful measures of harmoniousness. The presented results correlate significantly with empirical investigations on the perception of chords. Even for scales, plausible results are obtained. For example, all classical church modes appear in the front ranks of all theoretically possible seven-tone scales.