Job Openings

Bei der Technischen Universität Berlin ist/sind folgende Stelle/n zu besetzen: 

Wiss. Mitarbeiter/in - Entgeltgruppe 13 TV-L Berliner Hochschulen  (PhD candidate/PostDoc)

Counting people on image sequences using deep learning

Dauer: 2 Jahre

Besetzbar ab: so schnell wie möglich

Bewerbungsfrist: Position ist solange offen, bis jemand gefunden ist

Fakultät IV - Institut für Softwaretechnik und Theoretische Informatik / FG Neuronale Informationsverarbeitung Kennziffer

Aufgabenbeschreibung:

Der erfolgreiche Kandidat soll an einem IBB-geförderten Projekt (mit der Interautomation Berlin) teilnehmen, dessen Ziel es ist, Passagieren anhand von Tiefenbildern zu zählen. Dazu sollen aktuelle Erkenntnisse aus der Forschung zu neuronalen Netzen und sequentiellen Modellen genutzt werden.

Erwartete Qualifikationen:

Erfolgreich abgeschlossenes wissenschaftliches Hochschulstudium (Master, Diplom oder Äquivalent) in Informatik, Computational Neuroscience oder einem anderen MINT-Fach; sehr gute Kenntnisse in maschinellem Lernen und der Entwicklung von Software; sehr gute Deutsch- und Englischkenntnisse

Bewerbungsverfahren:

Zur Wahrung der Chancengleichheit zwischen Frauen und Männern sind Bewerbungen von Frauen mit der jeweiligen Qualifikation ausdrücklich erwünscht. Schwerbehinderte werden bei gleicher Eignung bevorzugt berücksichtigt. Die TU Berlin schätzt die Vielfalt Ihrer Mitglieder und verfolgt die Ziele der Chancengleichheit.

Aus Kostengründen werden die Bewerbungsunterlagen nicht zurückgesandt. Bitte reichen Sie nur Kopien ein.

Technische Universität Berlin offers an open position:

Research Assistant - salary grade E 13 TV-L (PhD candidate/PostDoc)

Risk-sensitive choice and reinforcement learning under uncertainty

Duration: 3 years

Starting Date: April, 1st, 2019

Application deadline: February, 28th, 2019

Faculty IV - Institute of Software Engineering and Theoretical Computer Science / Neural Information Processing

Working field:

The successful candidate is supposed to join a DFG funded project (with Co-PI Dirk Ostwald, Freie Universität Berlin) which combines computational modelling with behavioral and fMRI experiments. The goal is to better understand human decision making and reward-based learning under perceptual uncertainty. The computational framework generalizes the risk-sensitive MDP approach of Yun et al. 2014 (Neural Comput. 26, 1298ff) to a POMDP-based behavioral modelling framework. Planned experiments address response time behavior and neural reinforcement learning processes under perceptual risk. The work program can be biased either towards computational or experimental work depending on the qualifications of the candidate.

Qualifications:

Successfully completed university degree in a subject relevant to the work program; very good programming and English language skills; solid background in decision making and reward-based learning on a theoretical and / or experimental level; experience with computational modelling and the model-based analysis of behavioral and fMRI data is desirable

How to apply:

Please send your application with the usual documents (CV, letter of motivation, transcripts of records, certificates, and the names of two persons who can provide recommendation letters) to Prof. Dr. Klaus Obermayer, FG Neuronale Informationsverarbeitung, Sekr. MAR 5-6, Marchstr. 23, 10587 Berlin or by e-mail to klaus.obermayer@tu-berlin.de.

To guarantee equal opportunities between men and women, applications of women with respective qualifications are explicitly welcome. Severely handicapped persons will be privileged on equal qualification.

Please send copies only. Original documents will not be returned.

Technische Universität Berlin offers an open position:

Student Reseach Assistant (40h/Month)

Efficient Multi-task Deep Learning

Duration: 3 years

Starting date: August, 1st, 2019

Application deadline: July, 2nd, 2019

Aplication reference number: 3435 T 50/19

Faculty IV - Institute of Software Engineering and Theoretical Computer Science / Neural Information Processing

Working field:

The successful candidate will carry out experiments for training deep learning models and will analyze the generated representations for a battery of visual tasks. Various transfer learning paradigms will be implemented. Deep representations will be compared with visual representations in the human brain recorded during object recognition tasks. Efficient transfer learning techniques will be evaluated in the context of a synthetic system for rapid object recognition and visual search.

The position is part of the new DFG-funded cross-disciplinary research Cluster “Science of Intelligence” (http://www.scienceofintelligence.de/) and the successful candidate is welcome to participate in all of the Cluster's research and educational activities.

Qualifications:

Applicants should have very good programming skills, a good command of the English language, competence in machine learning, and a strong interest in working at the interface of machine learning and cognitive science. Some practical experience with deep learning techniques is a plus.

How to apply:

Please send your application with the usual documents (CV, letter of motivation, transcripts of records, and certificates) to Prof. Dr. Klaus Obermayer, Sekr. MAR 5-6, Marchstr. 23, 10587 Berlin, or (preferably) by e-mail to klaus.obermayer@tu-berlin.de.

To guarantee equal opportunities between men and women, applications of women with respective qualifications are explicitly welcome. Severely handicapped persons will be privileged on equal qualification.

Please send copies only. Original documents will not be returned.

Technische Universität Berlin offers an open position:

Student Research Assistant (60h/month)

Deep Networks for Modeling Neural Real-world Visual Categorical Decisions in Humans

Duration: 3 years

Starting Date: May 1st, 2019

Application deadline: February 15th, 2019

Faculty IV - Institute of Software Engineering and Theoretical Computer Science / Neural Information Processing

Working field:

The successful candidate will simulate deep networks for visual categorization tasks (including the piloting of alternative deep learning architectures) and evaluate these networks w.r.t. classification performance and the emerging visual representations. These networks will then be used (1) to calibrate visual datasets for experiments with human subjects and (2) to subsequently analyze the visual representations in the human brain recorded by MEG and fMRI.

The position is part of a DFG-funded project together with R. Cichy (FU Berlin).

Qualifications:

Applicants should have very good programming skills, a good command of the English language, competence in machine learning, and a strong interest in working at the interface of machine learning and cognitive neuroscience. Some practical experience with deep learning techniques is a plus.

How to apply:

Please send your application with the usual documents (CV, letter of motivation, transcripts of records, and certificates) to Prof. Dr. Klaus Obermayer, FG Neuronale Informationsverarbeitung, Sekr. MAR 5-6, Marchstr. 23, 10587 Berlin or by e-mail to klaus.obermayer@tu-berlin.de.

To guarantee equal opportunities between men and women, applications of women with respective qualifications are explicitly welcome. Severely handicapped persons will be privileged on equal qualification.

Please send copies only. Original documents will not be returned.

Technische Universität Berlin offers an open position:

Student Research Assistant (40h/month)

Reinforcement learning under risk

Duration: 3 years

Starting Date: As soon as possible (presumably 1.9.2019)

Application deadline: July 31th, 2019

Faculty IV - Institute of Software Engineering and Theoretical Computer Science / Neural Information Processing

Working field:

The successful candidate will work in the DFG-funded resarch project "Reinforcement learning under risk". The work focusses on mathematical analysis and numerical evaluation of risk-sensitive decision strategies.

Qualifications:

Good programming skills, good knowledge of mathematical theorie (propabilistic theorie and stochstics), very good english skills. Some knowledge of reinforcement learning is appreciated.

How to apply:

Please send your application with the usual documents (CV, letter of motivation, transcripts of records, and certificates) to Prof. Dr. Klaus Obermayer, FG Neuronale Informationsverarbeitung, Sekr. MAR 5-6, Marchstr. 23, 10587 Berlin or by e-mail to klaus.obermayer@tu-berlin.de.

To guarantee equal opportunities between men and women, applications of women with respective qualifications are explicitly welcome. Severely handicapped persons will be privileged on equal qualification.

Please send copies only. Original documents will not be returned.