@InProceedings{petrov-klein:2008:NIPS2008,
  author    = {Slav Petrov and Dan Klein},
  title     = {Discriminative Log-Linear Grammars with Latent Variables},
  booktitle = {Advances in Neural Information Processing Systems 20 (NIPS)},
  editor    = {J.C. Platt and D. Koller and Y. Singer and S. Roweis},
  publisher = {MIT Press},
  address   = {Cambridge, MA},
  pages     = {1153--1160},
  year      = {2008},
  url       = {http://books.nips.cc/papers/files/nips20/NIPS2007_0630.pdf}
}

We demonstrate that log-linear grammars with latent variables can be
practically trained using discriminative methods.  Central to
efficient discriminative training is a hierarchical pruning procedure
which allows feature expectations to be efficiently approximated  
in a gradient-based procedure. We compare L$_1$ and L$_2$ regularization
and show that L$_1$ regularization is superior, requiring fewer iterations 
to converge,
and yielding sparser solutions. On full-scale treebank parsing experiments, the
discriminative latent models outperform both the comparable generative
latent models as well as the discriminative non-latent baselines.

@InProceedings{petrov-pauls-klein:2007:EMNLP-CoNLL2007,
  author    = {Petrov, Slav  and  Pauls, Adam  and  Klein, Dan},
  title     = {Learning Structured Models for Phone Recognition},
  booktitle = {Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL)},
  pages     = {897--905},
  year      = {2007},
  url       = {http://www.aclweb.org/anthology/D/D07/D07-1094}
}

We present a maximally streamlined approach to learning HMM-based
acoustic models for automatic speech recognition.  In our approach, an
initial monophone HMM is iteratively refined using a
split-merge EM procedure which makes no assumptions about subphone
structure or context-dependent structure, and which uses only a single
Gaussian per HMM state.  Despite the much simplified training process,
our acoustic model achieves state-of-the-art results on phone
classification (where it outperforms almost all other methods) and
competitive performance on phone recognition (where it outperforms standard
CD triphone / subphone / GMM approaches). We also present an analysis
of what is and is not learned by our system.

@InProceedings{liang-EtAl:2007:EMNLP-CoNLL2007,
  author    = {Liang, Percy  and  Petrov, Slav  and  Jordan, Michael  and  Klein, Dan},
  title     = {The Infinite {PCFG} Using Hierarchical {Dirichlet} Processes},
  booktitle = {Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL)},
  pages     = {688--697},
  year      = {2007},
  url       = {http://www.aclweb.org/anthology/D/D07/D07-1072}
}

We present a nonparametric Bayesian model
of tree structures based on the hierarchical
Dirichlet process (HDP). Our HDP-PCFG
model allows the complexity of the grammar
to grow as more training data is available.
In addition to presenting a fully Bayesian
model for the PCFG, we also develop an efficient
variational inference procedure. On
synthetic data, we recover the correct grammar
without having to specify its complexity
in advance. We also show that our techniques
can be applied to full-scale parsing
applications by demonstrating its effectiveness
in learning state-split grammars.

@inproceedings{Petrov-Klein-2007:AAAI,
  author = {Slav Petrov and Dan Klein},
  title  = {Learning and Inference for Hierarchically Split {PCFG}s}
  booktitle = {AAAI 2007 (Nectar Track)},
  year = {2007},
  url = {www.eecs.berkeley.edu/~petrov/data/aaai2007.pdf},
}

Treebank parsing can be seen as the search for an optimally
refined grammar consistent with a coarse training treebank.
We describe a method in which a minimal grammar is hier-
archically refined using EM to give accurate, compact gram-
mars. The resulting grammars are extremely compact com-
pared to other high-performance parsers, yet the parser gives
the best published accuracies on several languages, as well
as the best generative parsing numbers in English. In addi-
tion, we give an associated coarse-to-fine inference scheme
which vastly improves inference time with no loss in test set
accuracy.

@InProceedings{petrov-klein:2007:main,

  author    = {Petrov, Slav  and  Klein, Dan},
    title     = {Improved Inference for Unlexicalized Parsing},
    booktitle = {Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics; Proceedings of the Main Conference},   
    month     = {April},
    year      = {2007},
    address   = {Rochester, New York},
    publisher = {Association for Computational Linguistics},
    pages     = {404--411},
    url       = {http://www.aclweb.org/anthology/N/N07/N07-1051}
}

We present several improvements to unlexicalized
parsing with hierarchically state-split PCFGs. First,
we present a novel coarse-to-fine method in which
a grammar's own hierarchical projections are used
for incremental pruning, including a method for efficiently
computing projections of a grammar without
a treebank. In our experiments, hierarchical
pruning greatly accelerates parsing with no loss in
empirical accuracy. Second, we compare various
inference procedures for state-split PCFGs from the
standpoint of risk minimization, paying particular
attention to their practical tradeoffs. Finally, we
present multilingual experiments which show that
parsing with hierarchical state-splitting is fast and
accurate in multiple languages and domains, even
without any language-specific tuning.

@inproceedings{Petrov-EtAl:2006:TRECVID,
  author = {Slav Petrov and Arlo Faria and Pascal Michaillat and Alexander Berg and Andreas Stolcke and Dan Klein and Jitendra Malik},
  title  = {Detecting Categories in News Video using Acoustic, Speech and Image Features},
  booktitle = {Proceedings of (VIDEO) TREC (TrecVid 2006)},
  year = {2006},
  url = {www.eecs.berkeley.edu/~petrov/data/trecvid06.pdf},
}

This work describes systems for detecting semantic categories 
present in news video. The multimedia data was processed in 
three ways: the audio signal was converted to a sequence of 
acoustic features, automatic speech recognition provided a 
word-level transcription, and image features were computed for
selected frames of the video signal. Primary acoustic, speech,
and vision systems were trained to discriminate instances of 
the categories. Higher-level systems exploited correlations 
among the categories, incorporated sequential context, and 
combined the joint evidence from the three information sources. 
We present experimental results from the TREC video retrieval 
evaluation.

@InProceedings{petrov-EtAl:2006:COLACL,
   author    = {Petrov, Slav  and  Barrett, Leon  and  Thibaux, Romain  and  Klein, Dan},
   title     = {Learning Accurate, Compact, and Interpretable Tree Annotation},
   booktitle = {Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics},
   month     = {July},
   year      = {2006},
   address   = {Sydney, Australia},
   publisher = {Association for Computational Linguistics},
   pages     = {433--440},
   url       = {http://www.aclweb.org/anthology/P/P06/P06-1055}
}

We present an automatic approach to tree annotation
in which basic nonterminal symbols are alternately
split and merged to maximize the likelihood
of a training treebank. Starting with a simple Xbar
grammar, we learn a new grammar whose nonterminals
are subsymbols of the original nonterminals.
In contrast with previous work, we are able
to split various terminals to different degrees, as appropriate
to the actual complexity in the data. Our
grammars automatically learn the kinds of linguistic
distinctions exhibited in previous work on manual
tree annotation. On the other hand, our grammars
are much more compact and substantially more accurate
than previous work on automatic annotation.
Despite its simplicity, our best grammar achieves
an F1 of 89.9% on the Penn Treebank, higher than
most fully lexicalized systems.

@InProceedings{petrov-barrett-klein:2006:CoNLL-X,
   author    = {Petrov, Slav  and  Barrett, Leon  and  Klein, Dan},
   title     = {Non-Local Modeling with a Mixture of {PCFG}s},
   booktitle = {Proceedings of the Tenth Conference on Computational Natural Language Learning (CoNLL-X)},
   month     = {June},
   year      = {2006},
   address   = {New York City},
   publisher = {Association for Computational Linguistics},
   pages     = {14--20},
   url       = {http://www.aclweb.org/anthology/W/W06/W06-2903}
}

While most work on parsing with PCFGs
has focused on local correlations between
tree configurations, we attempt to model
non-local correlations using a finite mixture
of PCFGs. A mixture grammar fit
with the EM algorithm shows improvement
over a single PCFG, both in parsing
accuracy and in test data likelihood. We
argue that this improvement comes from
the learning of specialized grammars that
capture non-local correlations.

@inproceedings{Tomasi-Petrov-Sastry-2003:ICCV,
  author = {Carlo Tomasi and Slav Petrov and Arvind Sastry},
  title  = {3{D} Tracking = {C}lassification + {I}nterpolation},
  booktitle = {Proceedings of the Ninth IEEE International Conference on Computer Vision (ICCV)},
  year = {2003},
  url = {www.eecs.berkeley.edu/~petrov/data/iccv03.pdf},
}

Hand gestures are examples of fast and complex motions.
Computers fail to track these in fast video, but sleight of
hand fools humans as well: what happens too quickly we
just cannot see. We show a 3D tracker for these types of
motions that relies on the recognition of familiar configurations
in 2D images (classification), and fills the gaps
in-between (interpolation). We illustrate this idea with experiments
on hand motions similar to finger spelling. The
penalty for a recognition failure is often small: if two con-
figurations are confused, they are often similar to each
other, and the illusion works well enough, for instance, to
drive a graphics animation of the moving hand. We contribute
advances in both feature design and classifier training:
our image features are invariant to image scale, translation,
and rotation, and we propose a classification method
that combines VQPCA with discrimination trees.

@mastersthesis{Petrov-Masters,
  author = {Slav Petrov},
  title  = {Computer vision, sensor fusion, and behavior control for soccer playing robots},
  school = {Freie Universitaet Berlin}
  year = {2004},
  url = {www.eecs.berkeley.edu/~petrov/data/slav_diplom_arbeit.pdf},
}

This Master's thesis describes parts of the control software
used by the soccer robots of the Free University of Berlin,
the so called FU-Fighters. The FU-Fighters compete in the
Middle Sized League of RoboCup and reached the semi-finals
during the 2004 RoboCup World Cup in Lisbon, Portugal. The
thesis covers several independent topics:
- Automatic White Balance: It is shown how to improve the
white balancing of an omni-directional camera by using a
reference color and a PID-controller.
- Ball Tracking: The reliable tracking of the ball is vital
in robot soccer. Therefore a Kalman-filter based system for
estimating the ball position and velocity in the presence
of occlusions is developped.
- Sensor Fusion: The robot perceives its environment through
several independent sensors (camera, odometer, etc.), which
have different delays. We propose a novel method for fusing
the sensor data and show our results through examples of
selflocalization.
- Behavior Control: Finally we show how all these elements
can be incorporated into a goal keeping robot. We develop
simple behaviors that can be used in a layered architecture
and enable the robot to block most balls that are being shot
at the goal.

• Carribean Side
• Montezuma
• Northern Pacific
• Lake Laguna
• Hot Springs
• Vulcanos
• Waterfalls
• Car
• Animals