Fair use or not, if you could do it I would buy it :) Fine, forget conference papers. If you can demonstrate fast NER in multiple languages, across domains, with competitive precision/recall metrics, I will buy it. The rest of it is not particularly interesting to me because it's frankly not that hard.