We propose a novel model training method for reranking problems. In our proposed approach, named the round-robin duel discrimination (R2D2), model training is done so that all pairs of samples can be distinguished from each other. The loss function of R2D2 for a log-linear model is concave. Therefore we can easily find the global optimum by using a simple parameter estimation method such a gradient descent method. We also describe the relationships between the global conditional log-linear model (GCLM) and R2D2. R2D2 can be recognized as an expansion of GCLM. We evaluate R2D2 on an error correction language model for speech recognition. Our experimental results using the corpus of spontaneous Japanese show that R2D2 provides an accurate model with a high generalization ability.
Bibliographic reference. Oba, Takanobu / Hori, Takaaki / Nakamura, Atsushi (2010): "Round-robin discrimination model for reranking ASR hypotheses", In INTERSPEECH-2010, 2446-2449.