Lesk algorithm

rewrite the LESK algorithm from WIKI.

Evaluation

After reviewing the papers, it was established that the evaluation of text generation should be carried out in four directions.

Relevance

Relevance is the most intuitive way of describing the extent to which the claimed text resembles the original text.

Redundancy

Redundancy is used to judge whether the generator has removed the maximum amount of unnecessary information from the original text.

Informativeness

The third metric is informativeness, which reflects how much the generated text can improve compared to the reader’s existing knowledge base.

Importance

Importance is specific to each semantic unit and reflects the intuitive weight of the generator, indicating how important the product is to the reader.

evaluation rank differ from experts and amateurs

We found that readers from different backgrounds evaluated the generators differently at different output scales, most likely because they were thought to hold different standards.

情绪

我每一天都在思考活下去的理由,但确实除了父母亲以外也想不到什么理由.这样的生活状态要么是上天给我的考验,要么就是上天给的提示,我恨我自己,我恨我的诞生,我恨我的存在.死在异国他乡虽然听起来很惨烈,但或许也没有那么多人会注意到吧.哈哈.