4 ms·
I'm pretty sure you could do this in a standard LALR(1) grammar and give [peanut butter] precedence over [peanut] [butter].
by sam0x17 8y ago
I'm pretty sure you could do this in a standard LALR(1) grammar and give [peanut butter] precedence over [peanut] [butter].
- alttab 8y agoThis would require that you have an exhaustive list of priorities typed out in a grammar, for each language. Word embeddings is a more semi-supervised learning. There is no way grammars could cover all the cases in a scalable way.
- sam0x17 8y agoTrue, so maybe the best approach is to use machine learning to generate the exhaustive list of priorities based on labeled human speech. Then your end product is something we can understand and tweak, instead of a black-box neural network.
- alttab 8y agoYou'd have to human-label that speech. That already won't scale very well, and requires per-language annotation. Understanding and tweaking it should be done with hyper parameters, not semantic libraries.
- anticensor 8y agoYou would need to look ahead recursively as different word orders may be acceptable in a human language. Not all languages are like French.