FG Kommunikationstechnik
Refine
Year of publication
Document Type
- Conference Proceeding (180)
- Conference publication peer-reviewed (30)
- Part of a book (chapter) (24)
- Scientific journal article peer-reviewed (23)
- Scientific journal article not peer-reviewed (15)
- Patent (9)
- Book (publisher) (8)
- Report (8)
- Doctoral thesis (6)
- Book (5)
Way of publication
- Open Access (8)
Keywords
- machine learning (4)
- Künstliche Intelligenz (3)
- Anfragesysteme (2)
- Information Retrieval (2)
- Informationskodierung (2)
- Kognitive Systeme (2)
- Lineare Algebra (2)
- Machine learning (2)
- Music psychology (2)
- Non-destructive testing (2)
Institute
- FG Kommunikationstechnik (316)
- FG Kognitive Materialanalytik (46)
- FG Angewandte Medienwissenschaften (10)
- FG Datenbank- und Informationssysteme (10)
- FG Automatisierungstechnik (2)
- FG Medientechnik (2)
- FG Stochastik und ihre Anwendungen (2)
- FG ABWL, insbesondere Marketing (1)
- FG Industrielle Informationstechnik (1)
- FG Ingenieurmathematik und Numerik der Optimierung (1)
Um eine bidirektionale Verarbeitung von Sprache zu realisieren, ist
es von Vorteil den gleichen Sprachformalismus für das Parsen und die Generierung
zu verwenden. Für Minimalistische Grammatiken gibt es eine ausgeprägte
Literatur zu Parsern, aber noch keine Veröffentlichung zu Generatoren. Diese Arbeit
stellt einen ersten Generator vor, der mittels Minimalistischer Grammatiken
und λ-Ausdrücken, Sätze erzeugt. Der hier vorgestellte Generator kann entgegen
manchen anderen Generatoren, von anderen Grammatikformalismen, mit ε-Regeln
arbeiten.
We further promote the idea of quantum inspiration and propose to equip cognitive systems not only with tensors for data representation but also for operation representation. We argue that these are two sides of the same coin. For experimental symbolic algorithms we introduce a suitable testbed and give its proper specification. We formalize our method of behavioural control with tensor algebra and discuss its implementation for our physical testbed realization.
We googled various numeral words from 28 languages. Different approaches
for the description of the data were investigated. In all of the 28 languages,
the found frequency-magnitude dependence fits better to a power law than
to an exponential law. The result can be used to distinguish grammatically correct
from incorrect numerals based on the prediction of search results.
This study presents recent advancements in Upper Sorbian
Speech-to-Text (STT) technology. We provide an overview of the Sorbian
languages, the available speech and language resources, and the development
of an STT system based on a traditional approach, which includes acoustic,
pronunciation, and language modeling.
Due to the scarcity of resources for Sorbian languages, our approach leverages
sub-word and word-class modeling techniques. The word-class modeling is based
on Finite-State Transducer definitions, which are applicable to both offline text
parsing and integration into the decoding graph of the STT system. Word-class
parsing is performed on the speech corpus and utilized for language modeling with
complete words, sub-word units, or both. Additionally, the same definitions can
be applied to Named Entity Recognition during the post-processing of recognized
transcriptions.
This approach significantly reduces out-of-vocabulary words and enables greater
customization of the recognizer for domain-specific applications. The system was
implemented for the real-time transcription of church sermon broadcasts in Upper
Sorbian. The domain-specific system achieved performance comparable to
fine-tuned OpenAI Whisper models developed also by other initiatives while also
providing a resource-efficient solution with semantically tagged recognition results.
Messungen an Quantenobjekten haben eine logische Struktur. Diese schließt die klassische Logik ein, geht jedoch weit darüber hinaus. Das wesentliche Ziel dieses Lehrbuchs ist es, die mathematischen Werkzeuge der Quantenlogik zu vermitteln und technische Anwendungsmöglichkeiten aufzuzeigen. Die dafür erforderlichen mathematischen Sachverhalte werden anhand von Beispielen so erläutert und motiviert, dass sie für angehende Ingenieure und Informatiker verständlich sind. Die vorliegende Auflage wurde korrigiert und um zahlreiche Anwendungen, zusätzliche Beispiele und Beweise erweitert sowie um ein neues Beispiel zur Mustererkennung ergänzt.
Quantum-inspired representation and querying of semiotic object properties by cognitive agents
(2025)
Advanced intelligence often relies on the ability to build and use models. They allow the representation of knowledge, enable predictions, and serve to plan goal-directed actions. However, the successfull use of models requires a certain level of competence in interacting with models. This competence can be acquired by exploiting the separation of the cognitive agent from the objects represented in the model; because it allows the differentiation of individual semiotic levels. This includes the separation of self-sufficient and non self-sufficient object attributes. Non self-sufficient attributes that the agent assigned to potential target objects during the exploration phase form the starting point for the information requirement regarding the self-sufficient attributes of these objects. These attributes are required for target selection and the subsequent alignment of the model-based planning. This process is characterized by a series of queries to the environment model and thus shows a certain similarity to information retrieval (IR). Therefore, in this work we particularly investigate quantum-inspired IR-methods for representation and querying. We demonstrate that they are suitable for modelling cognitive effort and behavior, as they represent a realistic and universally applicable descriptive tool.
We further promote the idea of quantum inspiration and propose to equip cognitive systems not only with tensors for data representation but also for operation representation and querying. We discuss ideas from information retrieval and translate them into our method of behavioral control. We formalize the method with tensor algebra and discuss its extensions to learning. At the example of the problem of 𝑁-dimensional aiming, we visualize the transfer of our stated theoretical foundation to a class of practical test problems.
Reducing feature dimensionality in spectroscopic data is crucial for efficient analysis and classification. Using all available features for classification typically results in an unacceptably high runtime and poor accuracy. Popular feature extraction methods, such as principal component analysis (PCA), linear discriminant analysis (LDA), and autoencoders, reduce feature dimensionality by extracting latent features that can be challenging to interpret. To enable better human interpretation of the classification model, we avoid extraction methods and instead propose applying feature selection methods. In this work, we develop an innovative PCA-based feature selection method for spectroscopic data, providing an essential subset of the original features. As an important advantage, no prior knowledge about the characteristic signals of the respective target substance is required. In this proof-of-concept study, the proposed method is initially characterized using simulated Raman and infrared absorption datasets. From the top five PCA eigenvectors of spectroscopic data, we identify a set of three top peaks each at specific wavenumbers (features). The compact set of selected features is then used for classification tasks applying a decision tree. Based on two well-defined spectroscopic datasets, our study demonstrates that our new method of PCA-based peak finding outperforms selected other approaches with regard to interpretability and accuracy. For both investigated datasets, accuracies greater than 97% are achieved. Our approach shows large potential for accurate classification combined with interpretability in further scenarios involving spectroscopic datasets.