The topic of the present workshop is stochastic optimal control theory and its relations to machine learning and robotics, statistical mechanics, quantum theory and the theory of large deviations.

For many years, the deterministic control theory has dominated control applications in robotics and autonomous systems, mainly because of computational restrictions. Relatively recently, there have been several approaches to restate the stochastic optimal control computation as an inference problem and to obtain efficient solutions using approximate inference. In these control theories, concepts from classical mechanics and control theory (variational calculus and Hamilton-Jacobi equations) and stochastic processes and large deviations theory (Feynman-Kac formula) are intimately related. This approach provides novel insights for the design of efficient algorithms to efficiently compute optimal stochastic control solutions in robotics.

The Hamilton Jacobi equation also plays a crucial role in the computation of non-equilibrium large deviations. In recent developments, this large deviation theory has provided a rather general framework in which the macrobehavior of non-equilibrium systems can be studied. In addition, there is a connection between these control formulations and Nelsons stochastic mechanics, which aims to provide a particle interpretation of quantum mechanics.

This workshop brings together researchers from control theory, machine learning, physics and mathematics to explore these connections.

Topics:

• Stochastic optimal control theory
• Stochastic processes
• Non-equilibrium large deviations
• Stochastic quantum theories and quantum control
• Applications in robotics

Statistical Physics of Inference and Control Theory

Statistical Physics of Inference and Control Theory

Statistical Physics of Inference and Control Theory

The goal of this workshop is to bring together researchers from both industry and academia to share their experiences of implementing large-scale applications of online learning and online decision-making. A selection of example applications includes banner advertisement selection, news story selection, targeted email, and recommender systems. The workshop will focus on the scalability of current online methods to large-scale implementations that are of practical value to industry. Relevant methods include exploration/exploitation trade-offs (e.g. contextual bandits), large-scale gradient descent, parallelization, collaborative filtering, unsupervised feature learning and dimensionality reduction.

The aim of this workshop is to contribute to the cross-fertilization between the research in machine learning methods and their applications to systems biology (i.e., complex biological and medical questions) by bringing together method developers and experimentalists. We encourage submissions bringing forward methods for discovering complex structures (e.g. interaction networks, molecule structures) and methods supporting genome-wide data analysis.

Molecular biology and all the biomedical sciences are undergoing a true revolution as a result of the emergence and growing impact of a series of new disciplines/tools sharing the "-omics" suffix in their name. These include in particular genomics, transcriptomics, proteomics and metabolomics, devoted respectively to the examination of the entire systems of genes, transcripts, proteins and metabolites present in a given cell or tissue type.

The BMVC Student Workshop will take place on Friday 7th September 2012, the day after the main BMVC conference, in the same venue. This workshop has become a regular feature of BMVC (see some of the previous editions: 2011, 2010). It gives students in computer vision an opportunity to network and start collaborations at an early stage in their research career. The workshop will be single track containing both oral and poster presentations.

Registration for this workshop is free for all UK students and BMVC registered participants.

Students studying in the UK are invited to submit full-length high-quality papers of which the main author is a student. All papers will be reviewed and selected for either oral or poster presentation.

Until recently, research in Natural Language Processing (NLP) has focused predominantly on propositional aspects of meaning. For example, semantic role labeling, question answering or text mining tasks aim at extracting information of the type ``who does what, when and where''. However, understanding language involves also processing Extra-Propositional Aspects of Meaning (EPAM), such as factuality, uncertainty, or subjectivity, since the same propositional meaning can be presented in a diversity of statements. While some work on phenomena like subjectivity has been carried out in the context of sentiment processing, other phenomena like the detection of sarcasm have received less attention.

By proposing this workshop we aim at bringing together scientists working on EPAM from any area related to computational language learning and processing. By EPAM we understand aspects of meaning that cannot be captured with a propositional representation such as the output of semantic role labelers.

For instance, the meaning of the sentence in Example (1) can be represented with the proposition ADD(earthquake,further threats to the global economy), whereas representing the meaning of the sentences in Example (2) requires additional mechanisms, despite the fact that all sentences share a propositional meaning.

1. The earthquake adds further threats to the global economy.
2. • Does the earthquake add further threats to the global economy?
   • The earthquake adds further threats to the global economy, doesn't it?
   • The earthquake does not add further threats to the global economy.
   • The earthquake will never add further threats to the global economy.
   • The earthquake will probably add further threats to the global economy.
   • Who could (possibly) think the earthquake adds further threats to the global economy?
   • The earthquake might have added further threats to the global economy.
   • The last analysis show that the earthquake will add further threats to the global economy.
   • It is expected that the earthquake will add further threats to the global economy.
   • It has been denied that the earthquake adds further threats to the global economy.

Some of the sentences above could also be combined in a paragraph such as (3), which shows that the same event can be presented from different perspectives, at different points in time and with different extra-propositional meanings.

3. The main question 6 months ago was whether the earthquake would add further threats to the global economy. Some days after the earthquake the authorities were convinced that it would be possible to minimize the impact of the earthquake. Most economists didn't share this view and predicted a high economic impact of the earthquake. However, a recent study about the earthquake's effect has shown that, although the earthquake might have added further threats to the global economy, its negative impact can be controlled by applying the right measures.

While the area of EPAM comprises a broad range of phenomena, this workshop will focus mainly on the aspects related to modality understood in a general sense (modalities, hedging, certainty, factuality), negation, attitude, and irony/sarcasm. Since many of these phenomena cannot be adequately modeled without taking (discourse) context into account, the workshop also touches on discourse phenomena in so far as they relate to extra-propositional aspects of meaning.

Topic

This workshop concerns analysis and prediction of complex data such as objects, functions and structures. It aims to discuss various ways to extend machine learning and statistical inference to these data and especially to complex outputs prediction. A special attention will be paid to operator-valued kernels and tools for prediction in infinite dimensional space.

Context and motivation

Complex data occur in many fields such as bioinformatics, information retrieval, speech recognition, image reconstruction, econometrics, biomedical engineering. In this workshop, we will consider two kinds of data: functional data and object or structured data. Functional data refers to data collected under the form of sampled curves or surfaces (longitudinal studies, time series, images). Analysis of these data as samples of random functions rather that a collection of individual observations is called Functional Data Analysis (FDA). FDA involves statistics in infinite-dimensional spaces and is closely associated to operatorial statistics. Its main approaches include functional principal component analysis and functional regression. Many theoretical challenges remain open in FDA and attract an increasing number of researchers.

Besides functional data, object and structure data exhibit an explicit structure like trees, graphs or sequences. For instance, documents, molecules, social networks and again images can be easily encoded as objet structured data. For the two last decades, both machine learning and statistics communities have developed various approaches to take into account the structure of the data. FDA is currently being extended to Object Data Analysis which deals with samples of object data instead of curves while in machine learning, graphical probabilistic models as well as kernel methods have been proposed among other methods to represent and analyze such data.

However, most of the efforts have been concentrated so far on dealing with complex inputs. In this workshop, we would like to emphasize the problem of complex outputs prediction which is involved for instance in multi-task learning, structured classification and regression, and network inference. All these tasks share a common feature: they can be viewed as approximation of vector-valued functions instead of scalar-valued functions and in the most general case, the output space is an Hilbert space. A promising direction first developed in (Micchelli and Pontil, 2005) consists in working with Reproducing Kernel Hilbert Spaces with operator-valued kernels in order to get an appropriate framework for regularization. There is thus a strong link between recent works in machine learning about prediction of multiple or complex outputs and functional and operatorial statistics.

This workshop aims at bringing together researchers from both communities to  1) provide an overview of existing concepts and methods, 2) identify theoretical challenges in  and (3) discuss practical applications and new tasks. To achieve this goal, we intend to build up from the successful workshops organized in the machine community about structured prediction like:

• Open House on Multi-Task and Complex Outputs Learning, UCL, 2007,
  
• Constrained Optimization and Learning with Structured Outputs at ICML 2007,
• Structured Input - Structured Output at NIPS 2008,
• Kernels for multiple outputs and multi task learning at NIPS 2009,
• Tensor, kernels and machine learning at NIPS 2010,
• International Workshop on Functional and Operational Statistics (IWFOS) in 2008 and 2011
• and the SAMSI program on Analysis of Object data in 2010-11.

Object, functional and structured data: towards next generation kernel-based methods – ICML 2012 Workshop

Object, functional and structured data: towards next generation kernel-based methods – ICML 2012 Workshop

We are organizing a workshop of gesture and sign language recognition from 2D and 3D video data and still images. Last year's workshop on gesture recognition at CVPR 2011 was a big success, with over 300 participants. This new workshop is coupled with a gesture recognition challenge http://gesture.chalearn.org, offering the opportunity to work on a large database of videos of hand gestures recorded with KinectTM. The best entrants will present their work at the workshop.

The scope of the workshop is broader than that of the challenge since gestures originate from any body motion and there is a wide variety of application settings in gaming, marketing, computer interfaces, interpretation of sign language for the deaf, and video surveillance. We invited keynote speakers in diverse areas of gesture research, including sign language recognition, body posture analysis, action and activity recognition, image or video indexing and retrieval, and facial expression or emotion recognition. The workshop aims at gathering researchers from different application domains working on gesture recognition to share algorithms and techniques.

It is possible to register to the workshop only, not to the full conference, but the participants have to register via the CVPR 2012 website. The calls for paper and for demonstration competitions are closed.

Gesture Recognition Workshop

Gesture Recognition Workshop

Morphologically Rich Languages (MRLs) are languages in which grammatical relations such as Subject, Predicate, Object, etc., are indicated morphologically (e.g. through inflection) instead of positionally (as in, e.g. English), and the position of words and phrases in the sentence may vary substantially. The tight connection between the morphology of words and the grammatical relations between them, and the looser connection between the position and grouping of words to their syntactic roles, pose serious challenges for syntactic and semantic processing. Furthermore, since grammatical relations provide the interface to compositional semantics, morpho-syntactic phenomena may significantly complicate processing the syntax--semantics interface. In statistical parsing, which has been a cornerstone of research in NLP and had seen great advances due to the widespread availability of syntactically annotated corpora, English parsing performance has reached a high plateau in certain genres, which is however not always indicative of parsing performance in MRLs, dependency-based and constituency-based alike . Semantic processing of natural language has similarly seen much progress in recent years. However, as in parsing,  the bulk of the work has concentrated on English, and MRLs may present processing challenges that the community is as of yet unaware of, and which current semantic processing technologies may have difficulty coping with. These challenges may lurk in areas where parses may be used as input, such as semantic role labeling, distributional semantics, paraphrasing and textual entailments, or where inadequate pre-processing of morphological variation hurts parsing and semantic tasks alike.

This joint workshop aims to build upon the first and second SPMRL workshops (at NAACL-HLT 2010 and IWPT 2011, respectively) while extending the overall scope to include semantic processing where MRLs pose challenges for algorithms or models initially designed to process English. In particular, we seek to explore the use of newly available syntactically and/or semantically annotated corpora, or data sets for semantic evaluation that can contribute to our understanding of the difficulty that such phenomena pose. One goal of this workshop is to encourage cross-fertilization among researchers working on different languages and among those working on different levels of processing. Of particular interest is work addressing the lexical sparseness and out-of-vocabulary (OOV) issues that occur in both syntactic and semantic processing.