Hauptinhalt
Topinformationen
Mitarbeiterverzeichnis
Neural Information Processing (Lecture + Practice)
Prof. Dr. Pascal Nieters
Veranstaltungstyp: Vorlesung und Übung
TeilnehmerInnen:
Beschreibung:
In the course Neural Information Processing, we will explore the principles of how neural systems compute and process information, focusing on the foundational models that have shaped our understanding of biological and artificial neural systems. The course is designed for advanced Bachelor students (4th semester+) and Master students in Cognitive Science, ready to gain a deep understanding of neural computation.
We will begin by studying single neurons as computational units. We’ll examine the perceptron and simple spiking neuron models, the perceptron and tempotron learning rules, and explore how these models explain the information processing capabilities of individual neurons. We will also investigate new insights into dendritic information processing that might shape our view of computation in neural computation going forward, and connect our discussions on single neurons to big picture ideas such as the information bottleneck principle.
We will then expand our focus and study feedforward neural networks. We’ll analyze how these networks learn and generalize, studying supervised mechanisms such as backpropagation and unsupervised ideas from neuroscience: Hebbian plasticity. We will link the fundamentals of network learning to modern machine learning architectures and our knowledge of topographic maps and connectomes in brains.
By taking this course, you will develop a deeper understanding of neural computation, refining your ability to connect theoretical models with real-world phenomena in cognitive science, neuroscience, and machine learning. The course will have a lecture component as well as a lab component (regular problem sets).
Prerequisites: Successfully participating in the course requires the ability to program in Python or Julia. At least one introductory course in mathematics and one course in statistics and/or probability theory are recommended. We will introduce additional concepts, for example from information theory, as needed.
Selection Process: The course is limited to 50 participants. Applicants are required to submit a short (<300 Words) motivational statement that addresses their particular interest in the course. If needed, the selection process for all applicants that have submitted the motivational statement is a lottery.
UPDATE: please send motivational statements by email to pascal.nieters@uni-osnabrueck.de. Random selection, if necessary, will happen at the beginning of the summer term.
Erstes Treffen:
Montag, 14.04.2025 10:00 - 12:00, Ort: 93/E31
Ort: 93/E31
Semester: SoSe 2025
Zeiten:Mo. 10:00 - 12:00 (wöchentlich) - Lecture, Do. 14:00 - 18:00 (wöchentlich) - Practice and Tutorial (Lab)
Leistungsnachweis:
Veranstaltungsnummer:
8.31301
ECTS-Kreditpunkte:
8
Bereichseinordnung:
Veranstaltungen > Cognitive Science > Bachelor-Programm Veranstaltungen > Cognitive Science > Master-Programm Courses Open to Exchange Students > Human Sciences (e.g. Cognitive Science, Psychology)
Prof. Dr. Pascal Nieters
Veranstaltungstyp: Vorlesung und Übung
TeilnehmerInnen:
Beschreibung:
In the course Neural Information Processing, we will explore the principles of how neural systems compute and process information, focusing on the foundational models that have shaped our understanding of biological and artificial neural systems. The course is designed for advanced Bachelor students (4th semester+) and Master students in Cognitive Science, ready to gain a deep understanding of neural computation.
We will begin by studying single neurons as computational units. We’ll examine the perceptron and simple spiking neuron models, the perceptron and tempotron learning rules, and explore how these models explain the information processing capabilities of individual neurons. We will also investigate new insights into dendritic information processing that might shape our view of computation in neural computation going forward, and connect our discussions on single neurons to big picture ideas such as the information bottleneck principle.
We will then expand our focus and study feedforward neural networks. We’ll analyze how these networks learn and generalize, studying supervised mechanisms such as backpropagation and unsupervised ideas from neuroscience: Hebbian plasticity. We will link the fundamentals of network learning to modern machine learning architectures and our knowledge of topographic maps and connectomes in brains.
By taking this course, you will develop a deeper understanding of neural computation, refining your ability to connect theoretical models with real-world phenomena in cognitive science, neuroscience, and machine learning. The course will have a lecture component as well as a lab component (regular problem sets).
Prerequisites: Successfully participating in the course requires the ability to program in Python or Julia. At least one introductory course in mathematics and one course in statistics and/or probability theory are recommended. We will introduce additional concepts, for example from information theory, as needed.
Selection Process: The course is limited to 50 participants. Applicants are required to submit a short (<300 Words) motivational statement that addresses their particular interest in the course. If needed, the selection process for all applicants that have submitted the motivational statement is a lottery.
UPDATE: please send motivational statements by email to pascal.nieters@uni-osnabrueck.de. Random selection, if necessary, will happen at the beginning of the summer term.
Erstes Treffen:
Montag, 14.04.2025 10:00 - 12:00, Ort: 93/E31
Ort: 93/E31
Semester: SoSe 2025
Zeiten:Mo. 10:00 - 12:00 (wöchentlich) - Lecture, Do. 14:00 - 18:00 (wöchentlich) - Practice and Tutorial (Lab)
Leistungsnachweis:
Veranstaltungsnummer:
8.31301
ECTS-Kreditpunkte:
8
Bereichseinordnung:
Veranstaltungen > Cognitive Science > Bachelor-Programm Veranstaltungen > Cognitive Science > Master-Programm Courses Open to Exchange Students > Human Sciences (e.g. Cognitive Science, Psychology)
