| Vorlesung Advanced Topics in Machine Learning for Human Language Technology | Lecture Advanced Topics in Machine Learning for Human Language Technology | 0 | 2 |
| Übung Advanced Topics in Machine Learning for Human Language Technology | Excercise Advanced Topics in Machine Learning for Human Language Technology | 0 | 1 |
| Prüfung Advanced Topics in Machine Learning for Human Language Technology | Exam Advanced Topics in Machine Learning for Human Language Technology | 4 | 0 |
### 3. Studien- und Prüfungsleistungen
*[Erläuterungen zu Abschnitt 3](../Anleitungen_Hilfen/3_Leistungen.md)*
**(!) Teilnahmevoraussetzungen (DE)**
Keine.
**(!) Teilnahmevoraussetzungen (EN)**
None.
**(!) Prüfungsbedingungen (DE)**
Klausur (100 %). Voraussetzung für die Zulassung zur Prüfung ist das Bestehen von Hausaufgaben.
**(!) Prüfungsbedingungen (EN)**
Written Exam (100 %). Students must pass written homework to be admitted to the examination.
### 4. Informationen für das Modulhandbuch
*[Erläuterungen zu Abschnitt 4](../Anleitungen_Hilfen/4_Infos_Modulhandbuch.md)*
**(!) Empfohlene Voraussetzungen (DE)**
Fundamentals of automatic speech recognition and machine learning, Statistical Classification and Machine Learning or further lectures on Machine Learning and/or Human Language Technology....
**(!) Empfohlene Voraussetzungen (EN)**
Fundamentals of automatic speech recognition and machine learning, Statistical Classification and Machine Learning or further lectures on Machine Learning and/or Human Language Technology.
**(!) Lernziele (DE)**
Knowledge:
Today data-driven methods like machine learning and artificial neural networks (ANN) are widely used for speech and language processing, e.g. for automatic speech recognition (ASR) and machine translation. We will re-visit the evolution of these methods over the last 50 years
and will present a unifying view of their principles from a probabilistic perspective.
On successful completion of this module, students should be able to:
• describe the components and formalisms of machine learning problems in human language technology from a principled probabilistic perspective;
• state and interpret the optimization problems of machine learning for human language technology and its relation to performance evaluation.
Skills:
They should be able to:
• apply state-of-the-art machine learning models to human language technology components;
• solve the optimization problems underlying decoding in sequence classification based on state-of-the-art machine learning components and underlying models;
• should have acquired soft skills like acquiring, reproducing and discussing knowledge about current state-of-the-art machine learning models for human language technology in a cooperative environment.
Competences:
• understand the probabilistic interpretation of artificial neural network outputs;
• describe the relation between the task performance (e.g. word error rate in ASR) and the decision rule for generating the output sequence (e.g. Bayes decision rule);
• relate training criteria (like cross-entropy) to task performance;
• be able to model the dependencies between input and output sequences in sequence-to-sequence processing;
• know the synchronization mechanisms between input and output sequences (e.g. hidden Markov models, finite-state transducers, cross-attention);
• understand the role language modeling plays in the context of end-to-end models.
**(!) Lernziele (EN)**
Knowledge:
Today data-driven methods like machine learning and artificial neural networks (ANN) are widely used for speech and language processing, e.g. for automatic speech recognition (ASR) and machine translation. We will re-visit the evolution of these methods over the last 50 years
and will present a unifying view of their principles from a probabilistic perspective.
On successful completion of this module, students should be able to:
• describe the components and formalisms of machine learning problems in human language technology from a principled probabilistic perspective;
• state and interpret the optimization problems of machine learning for human language technology and its relation to performance evaluation.
Skills:
They should be able to:
• apply state-of-the-art machine learning models to human language technology components;
• solve the optimization problems underlying decoding in sequence classification based on state-of-the-art machine learning components and underlying models;
• should have acquired soft skills like acquiring, reproducing and discussing knowledge about current state-of-the-art machine learning models for human language technology in a cooperative environment.
Competences:
• understand the probabilistic interpretation of artificial neural network outputs;
• describe the relation between the task performance (e.g. word error rate in ASR) and the decision rule for generating the output sequence (e.g. Bayes decision rule);
• relate training criteria (like cross-entropy) to task performance;
• be able to model the dependencies between input and output sequences in sequence-to-sequence processing;
• know the synchronization mechanisms between input and output sequences (e.g. hidden Markov models, finite-state transducers, cross-attention);
• understand the role language modeling plays in the context of end-to-end models.
**(!) Inhalt (DE)**
• Part 1: Probabilistic foundations, Bayes decision theory, probabilistic interpretation of neural networks, training criteria.
• Part 2: Sequence processing and specific ANN structures (hidden markov models, finite-state transducers, cross-attention).
• Part 3: Deep Learning and HLT tasks (automatic speech recognition, language modelling, machine translation).
**(!) Inhalt (EN)**
• Part 1: Probabilistic foundations, Bayes decision theory, probabilistic interpretation of neural networks, training criteria.
• Part 2: Sequence processing and specific ANN structures (hidden markov models, finite-state transducers, cross-attention).
• Part 3: Deep Learning and HLT tasks (automatic speech recognition, language modelling, machine translation).
**(!) Literatur**
### 5a. SPOen, in denen das Modul verankert werden soll
*[Erläuterungen zu Abschnitt 5a/5b](../Anleitungen_Hilfen/5_Studiengänge.md)*
