Thema der Dissertation:
Individuality in the hive A Data-Driven Study of Individual and Collective Behavior Thema der Disputation:
Decoding the Language of Life: Machine Learning and Biological Systems
Individuality in the hive A Data-Driven Study of Individual and Collective Behavior Thema der Disputation:
Decoding the Language of Life: Machine Learning and Biological Systems
Abstract: Our understanding of biological systems is rapidly advancing through new methodological developments. In the first talk, I will demonstrate the role of machine learning in enhancing our understanding of these systems. I will explore how recent breakthroughs in autoregressive language models are being adapted to the fundamental languages of biology: DNA, RNA, and proteins. This talk will clarify the basics of autoregressive language models using deep learning and outline the methodological progress required to apply these models to DNA sequences.
In the subsequent talk, I will discuss how combining automated tracking technologies with machine learning techniques enhances our understanding of complex behaviors within social groups. Using honey bee colonies as a model system, I will describe how large-scale datasets of social interactions enable the analysis of how individual behaviors contribute to the emergence of a group's collective behavior. I will introduce a new descriptor, "network age," derived from the spectral decomposition of social interaction networks, which effectively predicts future behavior and task allocation of the individuals. It offers insights into the developmental pathways of bees and their contributions to colony function through detailed behavioral analysis over their lifetimes.
In the subsequent talk, I will discuss how combining automated tracking technologies with machine learning techniques enhances our understanding of complex behaviors within social groups. Using honey bee colonies as a model system, I will describe how large-scale datasets of social interactions enable the analysis of how individual behaviors contribute to the emergence of a group's collective behavior. I will introduce a new descriptor, "network age," derived from the spectral decomposition of social interaction networks, which effectively predicts future behavior and task allocation of the individuals. It offers insights into the developmental pathways of bees and their contributions to colony function through detailed behavioral analysis over their lifetimes.
Zeit & Ort
18.04.2024 | 15:00
Raum K40
(Fachbereich Mathematik und Informatik, Takustr. 9, 14195 Berlin)
&
WebEx