3 RA positions in honeybee computational neuroscience and robotics

Three postdoctoral positions are available at the Universities of
Sheffield (two posts) and Sussex (one post) as part of the EPSRC funded,
‘Green Brain’ project. This exciting new project will develop
computational neuroscience models of learning and decision-making in the
honeybee brain, and controllers based on these to run on an NVIDIA GPU
supercomputer controlling a flying robot in real time. Invertebrate
neuroscientists are continuing to demonstrate that despite their small
sized brains, insects, such as honeybees, have comparable cognitive
sophistication to those of larger-brained animals, including
vertebrates. Honeybees, in particular, have been demonstrated to be able
to manage speed-accuracy trade-offs in decision-making, exhibit positive
and negative-reinforcement learning, and transfer concepts such as
‘sameness’ and ‘difference’ across sensory modalities. This project is
intended to advance our understanding of the invertebrate brain by
computational neuroscience modelling, with the ultimate long-term goal
of achieving a complete brain model of an animal such as the honeybee.
To achieve this goal, modern GPU super-computing will be used to build
detailed models of brain function that can run in real time and can
interface with a flying robot to study its behaviour in an embodied
context. The work will be carried out in close collaboration with
honeybee experts in Toulouse. It is expected that the long-term goal of
a full brain model will not only represent a significant basic research
achievement, but also lead to breakthroughs in artificial intelligence,
control of autonomous agents and computational insights into cognitive
mechanisms in higher animals

The postdoctoral positions of the research associates on this project
are as follows:

1. Computational neuroscientist (Sussex): Your primary responsibilities
will be to further develop models of the honeybee olfactory system and
learning pathways, develop GPU modelling tools, and integrate your work
with the other research associates.
http://www.jobs.ac.uk/job/AFD100/research-fellow-in-computational-neuroscience/

2. Computational neuroscientist (Sheffield): Your primary
responsibilities will be to model the honeybee optic tubercle and visual
learning pathways, to investigate multi-modal integration and learning,
and to integrate your work with the other research associates
http://www.jobs.ac.uk/job/AFE369/research-associate-in-computational-neuroscience/

3. Roboticist (Sheffield): Your primary responsibility will be to
develop and maintain the GPU-supercomputer-controlled flying robot, and
integrate the work of the other research associates into the platform
http://www.jobs.ac.uk/job/AFE372/research-associate-in-robotics/

Successful candidates must hold a PhD or equivalent degree in a
quantitative science discipline. All posts require a keen interest in
computational neuroscience and the basis of learning and behaviour in
animals. We are looking for candidates with a strong mathematical,
computational and computational neuroscience background (posts 1 and 2)
and keen interest in robotics (post 3). Knowledge of the insect
olfactory system (post 1), visual system (post 2) and robotic
controllers (post 3) is desirable, but is not a requirement. All
positions require good programming skills and experience with GPU
computing would be a big plus. The positions will involve travel between
Sheffield and Sussex and occasionally to the collaborating experimental
bee researchers in Toulouse.

For informal inquiries about the positions, please contact Dr. James
Marshall, James.Marshall@shef.ac.uk or Dr. Thomas Nowotny,
t.nowotny@sussex.ac.uk.

Candidates interested in applying for the University of Sussex job
please apply through www.sussex.ac.uk/jobs. Candidates interested in the
posts at University of Sheffield please apply through
http://www.shef.ac.uk/jobs. If candidates are interested in several
posts please apply on both sites. Sheffield Refs: UOS005250, UOS005253
Sussex Ref: 816.

Please provide a CV with publication list, a brief (1 page) statement of
why you are interested in the position and about your future career
plans with your application form.

Salary range: starting at £28,401 and rising to £37,012 per annum,
according to post and experience

Expected start date: 1 December 2012

Closing date for applications: 14/17 October 2012

Interviews are anticipated for: 1 November 2012

For full details and how to apply see
www.shef.ac.uk/jobs
www.sussex.ac.uk/jobs

The Universities of Sheffield and of Sussex are committed to equality of
opportunity.

Several Machine Learning / Natural Language Processing Researcher Positions at NICTA (Canberra, Australia)

There are several positions available for PhD qualified researchers at the intersection of machine learning and natural language processing at the Canberra Lab of NICTA.

Duration: Up to 3 years
Salary: $AU95-110k TRP
Closing Date: 8 October
Details: http://www.seek.com.au/Job/researchers/in/act-act/23106802

Deadline Extension: Multi-Trade-offs in Machine Learning, NIPS-2012 workshop

CALL FOR ABSTRACTS AND OPEN PROBLEMS
Multi-Trade-offs in Machine Learning
NIPS-2012 Workshop, Lake Tahoe, Nevada, US
https://sites.google.com/site/multitradeoffs2012/
December 7 or 8, 2012
———————————————-
We invite submission of abstracts and open problems to Multi-Trade-offs in Machine Learning NIPS-2012 workshop.
IMPORTANT UPDATES:
Deadline extension: we are extending the deadline until October 16.
Student scholarships: we are grateful for receiving support from the PASCAL network and will provide a limited number of travel scholarships to students and post-docs. Detailed information will be published in a few days on the web site.
IMPORTANT DATES
Submission Deadline: October 16.
Notification of Acceptance: October 30.
More details are provided below.
———————————————–
Abstract
One of the main practical goals of machine learning is to identify relevant trade-offs in different problems, formalize, and solve them. We have already achieved fairly good progress in addressing individual trade-offs, such as model order selection or exploration-exploitation. In this workshop we would like to focus on problems that involve more than one trade-off simultaneously. We are interested both in practical problems where “multi-trade-offs” arise and in theoretical approaches to their solution. Obviously, many problems in life cannot be reduced to a single trade-off and it is highly important to improve our ability to address multiple trade-offs simultaneously. Below we provide several examples of situations, where multiple trade-offs arise simultaneously. The goal of the examples is to provide a starting point for a discussion, but they are not limiting the scope and any other multi-trade-off problem is welcome to be discussed at the workshop.
Multi-trade-offs arise naturally in interaction between multiple learning systems or when a learning system faces multiple tasks simultaneously; especially when the systems or tasks share common resources, such as CPU time, memory, sensors, robot body, and so on. For a concrete example, imagine a robot riding a bicycle and balancing a pole. Each task individually (cycling and pole balancing) can be modeled as a separate optimization problem, but their solutions has to be coordinated, since they share robot resources and robot body. More generally, each learning system or system component has its own internal trade-offs, which have to be balanced against the trade-offs of other systems, whereas shared resources introduce external trade-offs that enforce cooperation. The complexity of interaction can vary from independent systems sharing common resources to systems with various degrees of relation between their inputs and tasks. In multi-agent systems communication between the agents introduces an additional trade-off.
We are also interested in multi-trade-offs that arise within individual systems. For example, model order selection and computational complexity [1], or model order selection and exploration-exploitation [2]. For a specific example of this type of problems, imagine a system for real-time prediction of the location of a ball in table tennis. This system has to balance between at least three objectives that interact in a non-trivial manner: (1) complexity of the model of flight trajectory, (2) statistical reliability of the model, (3) computational requirements. Complex models can potentially provide better predictions, but can also lead to overfitting (trade-off between (1) and (2)) and are computationally more demanding. At the same time, there is also a trade-off between having fast crude predictions or slower, but more precise estimations (trade-off between (3) and (1)+(2)). Despite the complex nature of multi-trade-offs, there is still hope that they can be formulated as convex problems, at least in some situations [3].
References:
[1] Shai Shalev-Shwartz and Nathan Srebro. “SVM Optimization: Inverse Dependence on Training Set Size”, ICML, 2008.
[2] Yevgeny Seldin, Peter Auer, François Laviolette, John Shawe-Taylor, and Ronald Ortner. “PAC-Bayesian Analysis of Contextual Bandits”, NIPS, 2011.
[3] Andreas Argyriou, Theodoros Evgeniou and Massimiliano Pontil. Convex multi-task feature learning. Machine Learning, 2008, Volume 73, Number 3.
Call for Contributions
We invite submission of abstracts and open problems to the workshop. Abstracts and open problems should be at most 4 pages long in the NIPS format (appendices are allowed, but the organizers reserve the right to evaluate the submissions based on the first 4 pages only). Selected abstracts and open problems will be presented as talks or posters during the workshop. Submission instructions will be published soon.
IMPORTANT DATES
Submission Deadline: October 16.
Notification of Acceptance: October 30.
EVALUATION CRITERIA
• Theory and application-oriented contributions are equally welcome.
• All the submissions should indicate clearly at least two non-trivial trade-offs they are addressing.
• Submission of previously published work or work under review is allowed, in particular NIPS-2012 submissions. However, for oral presentations preference will be given to novel work or work that was not yet presented elsewhere (for example, recent journal publications or NIPS posters). All double submissions must be clearly declared as such!
Invited Speakers
Shai Shalev-Shwartz, The Hebrew University of Jerusalem
Jan Peters, Technicsche Universitaet Darmstadt and Max Planck Institute for Intelligent Systems
Csaba Szepesvari, University of Alberta
Organizers
Yevgeny Seldin, Max Planck Institute for Intelligent Systems and University College London
Guy Lever, University College London
John Shawe-Taylor, University College London
Koby Crammer, The Technion
Nicolò Cesa-Bianchi, Università degli Studi di Milano
François Laviolette, Université Laval (Québec)
Gábor Lugosi, Pompeu Fabra University
Peter Bartlett, UC Berkeley and Queensland University of Technology
Sponsors
We are grateful for receiving support from the PASCAL network.
If you would also like to sponsor this event, please, contact seldin@tuebingen.mpg.de.
Tentative Schedule
7:30 – 7:35 Opening remarks
7:35 – 8:20 Invited Talk
8:20 – 8:50 Two Contributed Talks
8:50 – 9:10 Break
9:10 – 9:55 Invited Talk
9:55 – 10:30 Open Problems Session
10:30 – 15:30 Break
15:30 – 16:15 Invited Talk
16:15 – 16:25 Break
16:25 – 17:00 Two Contributed Talks
17:00 – 18:30 Posters
18:30 – 19:00 Workshop Summary and Open Discussion

Reminder: CFC NIPS Workshop on Discrete Optimization in Machine Learning (DISCML)

===================================================================

Call for Contributions

4th Workshop on
Discrete Optimization in Machine Learning (DISCML):
Structure and Scalability

at the Annual Conference on Neural Information Processing Systems (NIPS 2012)

http://www.discml.cc

Submission Deadline: Sunday 16th September, 11:59pm Samoa time

===================================================================
– We apologize for multiple postings –

Optimization problems with ultimately discretely solutions are becoming increasingly important in machine learning: At the core of statistical machine learning is to infer conclusions from data, and when the variables underlying the data are discrete, both the tasks of inferring the model from data, as well as performing predictions using the estimated model are discrete optimization problems. Two factors complicate matters: first, many discrete problems are in the general case very hard, and second, machine learning applications often demand solving such problems at large scale. The focus of this year’s workshop lies on structures that enable scalability. Which properties of the problem make it possible to still efficiently obtain exact or decent approximate solutions? What are the challenges posed by parallel and distributed processing? Which discrete problems in machine learning are in need of more scalable algorithms? How can we make disrete algorithms scalable? Some heuristics perform well but are as yet devoid of a theoretical foundation. What explains this behaviour?

We would like to encourage high quality submissions of short papers relevant to the workshop topics. Accepted papers will be presented as spotlight talks and posters. Of particular interest are new algorithms with theoretical guarantees, as well as applications of discrete optimization to machine learning problems.

Areas of interest include

Optimization

• Combinatorial algorithms
• Submodular / supermodular optimization • Discrete Convex Analysis • Pseudo-boolean optimization • Parallel & distributed discrete optimization

Continuous relaxations

• Sparse approximation & compressive sensing • Regularization techniques • Structured sparsity models

Learning in discrete domains

• Online learning / bandit optimization
• Generalization in discrete learning problems • Adaptive / stochastic optimization

Applications

• Graphical model inference & structure learning • Clustering • Feature selection, active learning & experimental design • Structured prediction • Novel discrete optimization problems in ML, Computer Vision, NLP, …

Submission deadline: September 16, 2012

Length & Format: max. 6 pages NIPS 2012 format

Time & Location: December 7 or 8 2012, Lake Tahoe, Nevada, USA

Submission instructions: Email submit@discml.cc

Invited talks by

• Satoru Fujishige
• Amir Globerson
• Alex Smola

Organizers:
Andreas Krause (ETH Zurich, Switzerland), Jeff A. Bilmes (University of Washington), Pradeep Ravikumar (University of Texas, Austin), Stefanie Jegelka (UC Berkeley)

CFP: HiPot: Workshop on Higher-Order Models and Global Constraints in Computer Vision and TPAMI Special Issue

CALL FOR PARTICIPATION

HiPot: Workshop on Higher-Order Models and Global Constraints in Computer Vision
ECCV 2012, Firenze, Italy
October 13, 2012
http://ttic.edu/eccv-2012-workshop-hipot/

A number of problems in computer vision can be modelled as discrete labelling problems. One of the popular ways to formulate a labelling problem has been in terms of an energy function comprising of unary and pairwise clique potentials. More recently, a second wave of success can be attributed to the incorporation of higher-order terms that have the ability to encode significantly more sophisticated priors and structural dependencies between variables.

The goal of this workshop is to bring together researchers working on different aspects (modelling, learning, inference) of higher-order models and global constraints for vision problems.

The workshop will include keynote talks from:
Endre Boros, Rutgers University
Fredrik Kahl, Lund University
Nikos Komodakis, Ecole des Ponts-ParisTech
Yann LeCun, New York University
René Vidal, Johns Hopkins University.

The complete program is available at: http://ttic.edu/eccv-2012-workshop-hipot/.

———————————————————————
CALL FOR PAPERS : TPAMI Special Issue

There will be a special issue on “Higher Order Graphical Models in Computer Vision: Modelling, Inference & Learning” associated with the workshop, to be published in IEEE Transactions on Pattern Analysis and Machine Intelligence. The call for papers is available at: http://www.di.ens.fr/~alahari/tpami-cfp.html. The timeline is as follows:

Submission Deadline: April 1, 2013
Reviews: October 1, 2013
Revisions of Submissions: January 1, 2014
Final Decisions/Manuscript: April 1, 2014
Estimated Online Publication: Fall 2014

Looking forward to seeing you in Florence!

CFP: CVMP2012

CVMP2012
5th & 6th December 2012.
Vue Leicester Square, London, UK
=================
Call for Short Papers
=================
Deadline : Friday 19th October.
————————
If you are doing interesting and innovative work or research in visual media we’d really like to read about it.
High-quality papers are invited which present novel research related to any aspect of media production. CVMP is held in cooperation with ACM SIGGRAPH, ACM MM, and EG. Accepted papers will be presented as either oral or posters and will appear in the conference programme. The top 10% of submissions to CVMP 2012 will be invited to submit extended papers to IEEE Transactions on Multimedia.
visit: http://www.cvmp-conference.org/Call-for-Papers
We look forward to receiving your papers.
Jan Kautz, UCL
CVMP2012 Programme Chair

Second CfP: 35th European Conference on Information Retrieval (ECIR’13)

Second CALL FOR PAPERS/POSTERS/DEMOS
35th European Conference on Information Retrieval (ECIR 2013) Moscow, Russia, 24-27 March 2013 http://ecir2013.org/

Updates:
– Four weeks to go for full papers—please submit your abstract timely!
– Open for submissions at http://www.conftool.pro/ecir2013/

Important dates:
– 1 Oct 2012: full paper abstract deadline
– 8 Oct 2012: full paper deadline
– 22 Oct 2012: posters/demos deadline
– 30 Nov 2012: notification of acceptance

Call for Papers

The conference encourages the submission of high-quality research papers reporting original and innovative research in Information Retrieval.
Submissions will be reviewed by experts on the basis of the originality of the work, the validity of the results, chosen methodology, writing quality and the overall contribution to the field of Information Retrieval. We accept not only full-papers, but also poster and demo short papers. Posters should present work in progress or leading-edge work. Demo papers should describe first-hand experiences with research prototype systems.

ECIR has traditionally had a strong student focus, and papers whose sole or main author is a postgraduate student or postdoctoral researcher are especially welcome. Papers that demonstrate a high level of research adventure or which break out of the traditional IR paradigms are also particularly welcome.

All submissions must be written in English following the LNCS author guidelines and submitted electronically through the conference submission system. Full papers must not exceed 12 pages including references and figures. Poster and Demo papers must not be longer than
4 pages. Accepted papers will have to be presented at the conference.
Posters and demonstrations will be presented at a special posters and demonstrations session. All submissions will be refereed. Papers and posters will undergo double-blind peer review, so authors should take reasonable care not identify themselves in their submissions. Demo submissions are not anonymous and should preferably contain a link to an online demo.

Full papers, posters and demos submissions in PDF conforming to the LNCS style can be submitted at http://www.conftool.pro/ecir2013/

Topics

The Program Chairs invite for the submission of original research papers and posters in all areas of Information Retrieval, including but not limited to:

* IR Theory and Formal Models:
– Searching, browsing, meta-searching, data fusion, filtering and indexing
– Text and content classification, categorisation, clustering
– Relevance feedback, query expansion
– Topic detection and tracking, novelty detection
– Content-based filtering, collaborative filtering, Spam filtering
– Personalised, collaborative or user-adaptive IR, recommender systems
– Adversarial IR
– Privacy in IR
– Mobile, Geo and Local Search

* Web and Social Media IR:
– Link analysis
– Query log analysis
– Advertising and ad targeting
– Spam detection
– Authority, Reputation, Ranking
– Blog and online-community search
– Social Tagging

* User aspects:
– User modelling, user studies, user interaction in IR systems
– Interactive IR, User studies, User models, Task-based IR
– Novel user interfaces for IR systems
– User interfaces, visualisation and presentation of queries, search results or content
– Multimodal aspects

* IR system architectures
– Distributed and peer to peer IR
– Parallel IR
– Fusion/Combination
– Open, interoperable and flexible
– Performance, Scalability, Architectures, Efficiency, Platforms
– Compression, performance, optimisation

* Content representation and processing
– IR for semi-structured documents
– IR for semantically annotated collections, semantic search
– Meta information and structures, metadata
– Query representation, Query reformulation
– Text Categorisation and clustering
– Text data mining
– Opinion mining
– Cross-language retrieval, Multilingual retrieval
– Machine translation for IR
– Question answering, Natural language processing for IR, Summarization

* Evaluation
– Evaluation methods and metrics
– Building test collections and metrics
– Experimental design
– Crowdsourcing for evaluation
– User-oriented and user-centred test and evaluation

* Multimedia and cross-media IR
– Speech retrieval
– Image and video retrieval
– Digital music, radio and broadcast retrieval

* Applications
– Digital libraries
– Enterprise Search, Intranet search, Desktop search
– Mobile IR
– Genomic IR, IR for chemical structures, etc.
– Medical IR, legal IR, patent search

Program Committee co-chairs:
– Jaap Kamps (University of Amsterdam, the Netherlands)
– Stefan Rüger (The Open University, UK)

Poster Chair:
– Eugene Agichtein (Emory University, USA)

Demonstrations Chair:
– Emine Yilmaz (Microsoft Research Cambridge, UK)

Postdoc and PhD positions at the University of Edinburgh

Postdoc and PhD positions in machine learning for systems biology are available at the School of Informatics of the University of Edinburgh.

The project, funded by the European Research Council, aims at developing novel statistical methodologies to model stochastic behaviour in biological systems. The postdoc/ student will work in close collaboration with experimental groups at SynthSys, the University of Edinburgh’s world leading Centre for Systems and Synthetic biology. For further details, please contact Guido Sanguinetti (gsanguin@inf.ed.ac.uk) and the following links

http://www.jobs.ed.ac.uk/vacancies/index.cfm?fuseaction=vacancies.detail&vacancy_ref=3016211&go=GO
http://www.ed.ac.uk/schools-departments/informatics/postgraduate/fees/research-grant-funding/machinelearninginbiology

PostDoc position in Computer Vision at Xerox Research

Starting Date: January 2013 or later.
The Textual and Visual Pattern Analysis group of the Xerox Research Centre Europe, is looking for a PostDoc researcher on the topic of Fine-Grained Visual Categorization (FGVC).
FGVC consists in classifying an image into a possibly large set of categories which are visually similar and semantically related e.g. animals, plants or vehicles. This is a challenging problem as the difference between two categories might rely on subtle cues which are difficult to detect, even with state-of-the-art algorithms. Although some systems have been successfully developed for various special-purpose applications such as the recognition of leaves or birds, one of the major challenges is to develop algorithms which can be applied to new problems or new domains with as little customization as possible. This project is carried out in collaboration with the TEXMEX group at INRIA Rennes (and especially Hervé Jégou) in the framework of a project sponsored by the French Government.
Requirements:
• Ph.D. degree in computer vision and / or machine learning.
• Excellent publication record in major computer vision / machine learning / multimedia conferences and journals.
• Strong development skills in C/C++ and Matlab.
• Fluent in written and oral English.
• Highly motivated.
The Textual and Visual Pattern Analysis group specializes in understanding, organizing, retrieving and enhancing visual and hybrid content. We have extensive experience and state-of-the-art systems in image categorization, image retrieval, image enhancement, quality / aesthetic assessment and document image processing. Our technology has won numerous awards in public competitions such as the PASCAL Visual Object Challenge, ImageCLEF or the ImageNet Large Scale Visual Recognition Challenge.
The Xerox Research Centre Europe is located in Grenoble, in the heart of the French Alps, close to both the Italian and Swiss borders. Based in France, the centre is part of the global Xerox Innovation Group made up of 650 researchers and Engineers in five world renowned research and technology centers.

To submit an application, please send your CV, cover letter as well as the name of at least one reference to both xrce-candidates@xrce.xerox.com and florent.perronnin@xrce.xerox.com.

ECML PKDD 2013 will have a continuous journal submission track in addition to the regular conference submission. The journal submission track is now open.

ECML PKDD 2013
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Prague, Czech Republic, September 23 to 27.

First Call for Papers
The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD) provides an international forum for the discussion of the latest high-quality research results in all areas related to machine learning and knowledge discovery in databases, as well as their application in innovative application domains.

The 2013 edition of ECML PKDD will have, next to the usual proceedings track, a new journal track in collaboration with Machine Learning and Data Mining and Knowledge Discovery. Articles can be submitted to the journal track all year long, while the proceedings track will have one deadline in April.

Submissions to the journal track should meet the standards of the journals, and, additionally, be concise and lend themselves to oral presentation at a conference. Articles focusing on consolidation of earlier work are less suitable for this track. Submissions will benefit from a streamlined reviewing process that allows for notification within 8 weeks. Resubmission of revised versions is possible. Upon acceptance, submissions automatically earn a presentation slot at the conference, and an abstract of the article will be included in the proceedings. More information about this new submission model is available at www.ecmlpkdd2013.org.

Submissions are invited on all aspects of machine learning, knowledge discovery and data mining, including real-world applications. Journal submissions should present work that is novel, timely, and constitutes a clearly delineated piece of research that can be considered finished. Submissions to the proceedings track ideally present innovative ideas that are inspiring, provoke discussion, and/or are demonstrated to have a large potential.

Important criteria for all submissions are their:

* potential to inspire the research community by introducing new and relevant problems, concepts, solution strategies, and ideas
* contribution to solving a problem widely recognized as both challenging and important
* capability to address a novel area of impact of machine learning and data mining
* scientific rigor, correctness, reproducibility of experiments
* presentation quality: preciseness and clarity is required

IMPORTANT DATES :

Journal track:

* bi-weekly batch deadlines on Sundays (GMT). The 2012 deadlines are 12.8., 26.8., 9.9., 23.9., 7.10., 21.10., 4.11., 18.11., 2.12., and 16.12.
* currently open to submissions
* notification within 8 weeks (for submissions within the page restrictions)

Proceedings track:

* abstract submission: Thursday, April 18, 2013
* paper submission: Monday, April 22, 2013
* notification: Friday, June 14, 2013
* camera ready copy: Friday, June 28, 2013

Conference: September 23-27, 2013