In light of IBM supercomputer Watson’s Jeopardy! triumph – and geography blunder – last evening, a number of very interesting articles have been popping up about artificial intelligence (AI) systems.
One of them is Mark Liberman’s latest Language Log post encouraging native English speakers to help researchers improve automatic speech recognition (ASR). While AI systems are actually quite good at ASR these days, Liberman says, improvements can still be achieved in two ways: 1. by giving the system more training data and 2. through systematic error analysis.
This second approach relies heavily on human-generated transcripts, thus explaining why researchers are busy recruiting willing volunteers.
“[E]ven when ASR word error rate is similar to human inter-transcriber differences in percentage terms, the patterns of machine and human disagreements are often quite different. But in order to use a comparison of those patterns in error analysis, researchers need a large enough sample of human transcripts — that’s where you (may) come in!” writes Liberman.
If you have an hour to spare and would like to contribute to this ASR project, you can start by taking one of the following speech transcription tests:
|Test 1||Test 2|
|Test 3||Test 4|
For additional information, read more from Mark Liberman.