Discussions

• 1 You must be logged in to vote

  💬

  I feel that meticulous and careful manual annotation one by one is irreplaceable.

  guotong1988 started Nov 3, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  The essential result of GPT pre-training is sentence similarity.

  guotong1988 started Sep 8, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  Reading paper can not only learn the knowledge but also learn the writing.

  guotong1988 started Sep 7, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  Everything Begins With The Problem

  guotong1988 started Sep 7, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  Transform Data Science To Software Engineering By Limited Human Annotation

  guotong1988 started Aug 29, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  For ChatGPT, human-feedback's goal is to fix the wrong data in policy-model's dataset. There is no essential difference between reinforce learning and supervised learning, here.

  guotong1988 started May 26, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  In reinforcement learning, there are no so-called 'reward', there are only 'positive and negative' collected data.

  guotong1988 started Feb 20, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  For ChatGPT，if you can collect all the possible pre-train data, then you can just remove the bad-feedback data from predictions for reward model. if you can not collect all the possible pre-train data, then you need to correct the bad-feedback data from predictions for reward model.

  guotong1988 started Feb 20, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  In ChatGPT, can we remove the reward-model part?

  guotong1988 started Feb 7, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💡

  不需要很多人工标注的ChatGPT + 可以推理的ChatGPT

  guotong1988 started Feb 14, 2023 in Ideas

  

  

  0
• 1 You must be logged in to vote

  💬

  准备好的prompt模版，适用于把知识图谱映射为自然句子

  guotong1988 started Feb 13, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  如果不能接触 有物理载体的Robot 研究，能否把APP看成和用户世界交互的agent。

  guotong1988 started Nov 26, 2022 in General

  

  

  5
• 1 You must be logged in to vote

  💬

  model generalization = find the most similar data in training dataset

  guotong1988 started Dec 11, 2022 in General

  

  

  1
• 1 You must be logged in to vote

  💬

  In ChatGPT, what is the difference between Reinforcement-Learning-from-Human-Feedback and Data-Re-Label-from-Reward?

  guotong1988 started Jan 10, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  In a broad sense, all the machine learning tasks be viewed as classification.

  guotong1988 started Jan 10, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  Deep learning now is solving the problem of AI-agent remembering.

  guotong1988 started Jan 10, 2023 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  Model-centric = weighted sum of parameter

  guotong1988 started Dec 29, 2022 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  Data-centric = weighted sum of data

  guotong1988 started Dec 11, 2022 in General

  

  

  0
• 1 You must be logged in to vote

  💬

  如果能够开发一个足够真实的"模拟人生"，是否就可以实现更高阶的AI？

  guotong1988 started Oct 18, 2022 in General

  

  

  0
• 1 You must be logged in to vote

  📣

  Welcome to AI_thoughts Discussions!

  guotong1988 announced Oct 17, 2022 in Announcements

  

  

  0