I got a new idea for automatic summary generation task, but I was stuck in the middle.
I'm fighting for my new idea this week. It is about abstractive summary generation. You know, I'm a computational linguist.
Gigaword is the most commonly used dataset for the evaluation on this task. BUT, the annotated version needs 6000 US dollars. Yes, it is 6000 dollars!!! Can you believe that?
So what I decided to do, is to annotate it by myself. I spent a week processing the data, but I finally failed. The Stanford parser (a very famous natural language parser) always gave errors! (gave errors, is this expression correct?) "Your memory is not enough", "Sorry, the sentence is toooo long", "IO error",..., etc.
So, with no other choices, I give up this dataset. I won't touch it, at least, this month because I believe I will not get any result on it in a very short time. I must prove my thought quickly so and I'm trying to find another dataset.
Doing research is full of fun, but preprocess data is a boring job. XD
Wish me good luck in the next week! haha
Çeşitli İngilizce öğretmenleri arasından arama yapın...
Düzeltmeler · 2
0
Shake my head...\(-_-)/
0
被finally failed 击中笑点,这剧情走向不对啊
Daha hızlı mı ilerlemek istiyorsun?
Bu öğrenme topluluğuna katıl ve ücretsiz alıştırmaları dene!
Evden çıkmadan dil öğrenme fırsatını kaçırmayın. Deneyimli dil eğitmenlerimizden oluşan seçkimize göz atın ve ilk dersinize şimdi kaydolun!
Bugün öğrenmeye başlayın
