FG Kommunikationstechnik
Refine
Year of publication
Document Type
- Conference Proceeding (105)
- Conference publication peer-reviewed (19)
- Scientific journal article peer-reviewed (13)
- Part of a book (chapter) (10)
- Patent (7)
- Report (7)
- Scientific journal article not peer-reviewed (5)
- Image (poster) (3)
- Book (2)
- Doctoral thesis (1)
Way of publication
- Open Access (7)
Keywords
- machine learning (3)
- Machine learning (2)
- Non-destructive testing (2)
- Artificial intelligence (1)
- Cognitive system (1)
- Cognitive systems (1)
- Computational cognition (1)
- Computational linguistics (1)
- Data augmentation (1)
- Deep learning (1)
Institute
We present a Matlab toolbox, called “FockBox”, handling
Fock spaces and objects associated with Fock spaces: scalars, ket and
bra vectors, and linear operators. We give brief application examples
from computational linguistics, semantics processing, and quantum logic,demonstrating the use of the toolbox.
Quo vadis, UASR?
(2013)
We present a model, inspired by quantum field theory, of the so-called inner stage of technical cognitive agents. The inner stage represents all knowledge of the agent. It allows for planning of actions and for higher cognitive functions like coping and fantasy. By the example of a cognitive mouse agent living in a maze wold, we discuss learning, action planning, and attention in a fully deterministic setting and assuming a totally observable world. We explain the relevance of our approach to cognitive infocommunications.
In diesem Positionspapier stellen wir ausgehend von klassischen Erkenntnissen der Sprach– und Signalverarbeitung das im letzten Jahrzehnt von uns entwickelte "Einheitliche System zur Sprachsynthese und –erkennung" (UASR) vor. Dieses steht in enger Beziehung zu neueren Ansätzen in der Systemtheorie, namentlich den kognitiven dynamischen Systemen. Diese lassen bislang jedoch weitestgehend die für eine "Erkenntnisfähigkeit" unerlässliche hierarchische Modellierung und Systemstruktur außer acht (obwohl deren Notwendigkeit als unumstritten gelten kann). Die Konstruktionsprinzipien unseres Systems sind: hierarchische Struktur, gemeinsame Daten für Analyse und Synthese sowie einheitliche Algorithmen auf allen
Ebenen. Wir argumentieren weiterhin, dass kognitive (Sprach–)Kommunikationsysteme ein inneres Modell ihres Kommunikationspartners haben müssen und zeigen, wie dieses auf naheliegende Weise technisch realisiert werden kann. Sowohl die hierarchische Struktur als auch die ”Spiegelung“ des Kommunikationspartners sind durch die Neurobiologie beim natürlichen Vorbild belegt. Als technologische Basis schlagen wir endliche Transduktoren (finite state transducers) vor. Wir haben gezeigt, dass diese für unsere Aufgabe universell sind und prinzipiell auf allen Verarbeitungsebenen von der akustischen bis hin zur
pragmatischen für Analyse und Synthese eingesetzt werden
können. Abschließend geben wir ein Konzept für hierarchische kognitive dynamische Sprach- und Signalverarbeitungssysteme an und benennen die aus unserer Sicht künftig interessanten Forschungsthemen.
Akustische Mustererkennung
(2011)
A circle of a sphere is a circle drawn on a spherical surface like, for instance, circles of latitude or longitude. Circles in arbitrary 3D positions can be drawn with TikZ very easily using a transformed coordinate system provided by the tikz-3dplot package (that is because TikZ can only draw circles on the xy-plane). However, automatically distinguishing the parts of the circle lying on the front and back sides of the sphere, e.g. by drawing a solid arc on the front side and a dashed one on the back side, is a somewhat tricky feat. The tikz-3dplot-circleofsphere package will perform that feat for you.
Einkanalige Störgeräuschunterdrückung zur Steigerung der Worterkennungsrate eines Spracherkenners
(2007)
Ultrasonic Testing (UT) has seen increasing application of machine learning (ML) in recent years, promoting higher-level automation and decision-making in flaw detection and classification. Building a generalized training dataset to apply ML in non-destructive evaluation (NDE), and thus UT, is exceptionally difficult since data on pristine and representative flawed specimens are needed. Yet, in most UT test cases flawed specimen data is inherently rare making data coverage the leading problem when applying ML. Common data augmentation (DA) strategies offer limited solutions as they don’t increase the dataset variance, which can lead to overfitting of the training data. The virtual defect method and the recent application of generative adversarial neural networks (GANs) in UT are sophisticated DA methods targeting to solve this problem. On the other hand, well-established research in modeling ultrasonic wave propagations allows for the generation of synthetic UT training data. In this context, we present a first thematic review to summarize the progress of the last decades on synthetic and augmented UT training data in NDE. Additionally, an overview of methods for synthetic UT data generation and augmentation is presented. Among numerical methods such as finite element, finite difference, and elastodynamic finite integration methods, semi-analytical methods such as general point source synthesis, superposition of Gaussian beams, and the pencil method as well as other UT modeling software are presented and discussed. Likewise, existing DA methods for one- and multidimensional UT data, feature space augmentation, and GANs for augmentation are presented and discussed. The paper closes with an in-detail discussion of the advantages and limitations of existing methods for both synthetic UT training data generation and DA of UT data to aid the decision-making of the reader for the application to specific test cases.
Akustische Mustererkennung
(2011)