INTRODUCTION

 

With recent advances in information technology, digital archiving is emerging as an important and practical method for capturing the human experience. We propose to generate a quantum leap in the ability to access the contents of these archives by advancing the state of the art in automated speech recognition (ASR) and other component technologies. Before archives can be used efficiently, their contents must first be described, through some combination of human effort and automation. The scale of the collections and the difficulty of the task make it impractical to accomplish this entirely manually, particularly in the multilingual environment that globally interconnected information systems produce. Automatic technologies for search and exploration in spoken materials presently have relatively limited capabilities; capabilities that must be dramatically enhanced if the full potential of digital archiving is to be realized. In this project, we seek to make just such a leap, utilizing the world’s largest digital archive of video oral histories.  The multimedia digital archive collected by the Survivors of the Shoah Visual History Foundation contains over 116,000 hours of interviews with over 52,000 survivors, liberators, rescuers and witnesses of the Nazi Holocaust, recorded in 32 languages. Four thousand of the interviews in English have been manually cataloged at great expense, producing an exceptional source of labeled training data. In this project, we will use the unique features of this collection – massive quantities of multilingual audio and an extensive set of labeled training data – to achieve critical breakthroughs that will dramatically improve access to the recorded heritage of human civilization in this collection and others.