Publications

@InProceedings{ Royer_2016_ICASSP,
  author = {Royer, Amelie, Gravier, Guillaume and Claveau, Vincent},
  title = {Audio word similarity for clustering with 
           zero resources based on iterative HMM classification},
  booktitle = {The IEEE International Conference on Acoustics, 
               Speech and Signal Processing (ICASSP)},
  month = {March},
  year = {2016}
} 

Recent work on zero resource word discovery makes intensive use of audio fragment clustering to find repeating speech patterns. In the absence of acoustic models, the clustering step traditionally relies on dynamic time warping (DTW) to compare two samples and thus suffers from the known limitations of this technique.

We propose a new sample comparison method, called 'similarity by iterative classification', that exploits the modeling capacities of hidden Markov models (HMM) with no supervision. The core idea relies on the use of HMMs trained on randomly labeled data and exploits the fact that similar samples are more likely to be classified together by a large number of random classifiers than dissimilar ones.

The resulting similarity measure is compared to DTW on two tasks, namely nearest neighbor retrieval and clustering, showing that the generalization capabilities of probabilistic machine learning significantly benefit to audio word comparison and overcome many of the limitations of DTW-based comparison.

@InProceedings{ Royer_2015_CVPR,
  author = {Royer, Amelie and Lampert, Christoph H.},
  title = {Classifier Adaptation at Prediction Time},
  booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  month = {June},
  year = {2015}
} 

Classifiers for object categorization are usually evaluated by their accuracy on a set of i.i.d. test examples. This provides us with an estimate of the expected error when applying the classifiers to a single new image. In real application, however, classifiers are rarely only used for a single image and then discarded. Instead, they are applied sequentially to many images, and these are typically not i.i.d. samples from a fixed data distribution, but they carry dependencies and their class distribution varies over time.

In this work, we argue that the phenomenon of correlated data at prediction time is not a nuisance, but a blessing in disguise. We describe a probabilistic method for adapting classifiers at prediction time without having to retraining them. We also introduce a framework for creating realistically distributed image sequences, which offers a way to benchmark classifier adaptation methods, such as the one we propose.

Experiments on the ILSVRC2010 and ILSVRC2012 datasets show that adapting object classification systems at prediction time can significantly reduce their error rate, even with additional human feedback.

@techreport{royer:hal-01285965,
  TITLE = {{Similarity by diverting supervised machine learning - Application to knowledge discovery in multimedia content}},
  AUTHOR = {Royer, Amelie and Claveau, Vincent and Gravier, Guillaume and Furon, Teddy},
  URL = {https://hal.inria.fr/hal-01285965},
  TYPE = {Research Report},
  NUMBER = {RR-8880},
  INSTITUTION = {{Inria Rennes Bretagne Atlantique ; UMR IRISA}},
  YEAR = {2015},
  HAL_ID = {hal-01285965},
  HAL_VERSION = {v1},
}

Knowledge discovery is the task of extracting new information from large databases, such as recurrent patterns or structural cues. In this framework, cluster analysis refers to the sub-domain dealing with partitioning a given data space such that two samples in the same cluster are similar, while those in different ones are not. Clustering algorithms exploit an input similarity measure on the samples, which should be fine-tuned with the data format and the application at hand.

However, manually defining a suitable similarity measure is a difficult task in case of limited prior knowledge or complex data structures for example.The purpose of this work is to investigate an approach for automatically building such a measure by taking advantage of the discriminative abilities of state-of-the-art classification techniques. While classification systems usually require a set of samples annotated with their ground-truth classes, recent work has shown it is possible to exploit classifiers trained on an artificial annotation of the data in order to induce a similarity measure.

In this report, after introducing related scientific background, we propose a unified framework, SIC (Similarity by Iterative Classification), which explores the idea of diverting supervised learning for automatic similarity inference. We study several of its theoretical and practical aspects. We also implement and evaluate SIC on three tasks of knowledge discovery on multimedia content. Results show that in most situations the proposed approach indeed benefits from the underlying classifier's properties and outperforms usual similarity measures for clustering and information retrieval applications.