I'm so lost. I'll just explain how I would solve it =).
input/output: The old dragon left a pool of water in the canyon and landed on a hill to eat some grass. Where was the water left? Answers: Pool/canyon Answer: Canyon Explanation: The keywords are where, water, left, because of their rare frequency in big data. Using ex. word2vec we discover where=place/canyon/cave/etc. So activations of 'where' etc are matched: The old [dragon] left a [pool] of [water] in the [canyon] and landed on a [hill] to eat some [grass]. The old dragon left a [pool] of [water] in the canyon and landed on a hill to eat some grass. The old dragon [left] a pool of water in the canyon and [landed] on a hill to eat some grass. It's already clear most weight is on the left side. We could have: The dragon had the water. The water was in a pool. Left in the cave was water. Water in the cave no water was in the house yup. You're really looking for a 'match', which doesn't like unrelated words, stretched out matches, backwards matches, etc. Depending on the training data. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T21bdc2c440c86db7-M7ee7367d9c2f292755a1b664 Delivery options: https://agi.topicbox.com/groups/agi/subscription