Learning to Navigate in Cities Without a Map 理解
問題定義 在真實世界中進行無定位輔助，類似於人直覺長距離導航。輸入爲當前的視覺輸入和目標地點。輸出就是接下來應該怎麼走，才能到達目的地。PS：Navigation相比於planning來說更加粗糙，就是不需要具體到某個地點，而是一個大概的方向確定。 問題解決 使用DRL，定義狀態：當前視覺輸入；定義動作：五個離散動作，快慢向前，這個應該是由Google Map採集地圖的攝像頭確定的；獎勵：最開始的
相關文章
- 1. LEARNING TO NAVIGATE IN COMPLEX ENVIRONMENTS 閱讀筆記
- 2. Learning to Navigate for Fine-grained Classification
- 3. Top 5 Programming Languages to Learn in 2020 to Get a Job Without a College Degree
- 4. How to Learn Programming Without an Internet?
- 5. Procedural Modeling of Cities
- 6. Use Wireshark to capture loopback traffic without a loopback adapter （轉）
- 7. 論文閱讀筆記 | (ECCV 2018) Learning to Navigate for Fine-grained Classification
- 8. A strategy to apply machine learning to small datasets in materials science
- 9. A road map to become a Data Scientist（上）
- 10. Assignment | 01-week1-Introduction to deep learning
- 更多相關文章...
- • Scala Map(映射) - Scala教程
- • SQL IN 操作符 - SQL 教程
- • Docker 清理命令
- • Java Agent入門實戰（三）-JVM Attach原理與使用
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. LEARNING TO NAVIGATE IN COMPLEX ENVIRONMENTS 閱讀筆記
- 2. Learning to Navigate for Fine-grained Classification
- 3. Top 5 Programming Languages to Learn in 2020 to Get a Job Without a College Degree
- 4. How to Learn Programming Without an Internet?
- 5. Procedural Modeling of Cities
- 6. Use Wireshark to capture loopback traffic without a loopback adapter （轉）
- 7. 論文閱讀筆記 | (ECCV 2018) Learning to Navigate for Fine-grained Classification
- 8. A strategy to apply machine learning to small datasets in materials science
- 9. A road map to become a Data Scientist（上）
- 10. Assignment | 01-week1-Introduction to deep learning