reinforcement
US /ˌriɪnˈfɔrsmənt, -ˈfors-/
・UK /ˌri:ɪn'fɔ:smənt/
B2 中高級多益
n.名詞增援；加強
The commander sent a reinforcement of 1000 men to the soldiers already there
n.名詞加強；鞏固
My husband gave the reinforcement I needed to finish my paper
n.名詞加強；鞏固
The bridge needed reinforcement to withstand the increased traffic.
n.名詞增援部隊；增援物資
The troops received reinforcements just in time to defend the city.
n.名詞強化；增強 (心理學)
Positive reinforcement is more effective than punishment in training animals.
n.名詞加固材料；鋼筋
The concrete contained steel reinforcement to increase its tensile strength.
影片字幕
ChatGPT 是如何讓 OpenAI 成為最有價值的 AI 新創公司？ (How ChatGPT Made OpenAI One Of The Most Valuable AI Startups)
- The company released its first project in 2016, a toolkit called OpenAI Gym, used for developing and comparing reinforcement learning algorithms.
該公司於 2016 年發佈了第一個項目，一個名為 OpenAI Gym 的工具包，用於開發和比較強化學習算法。
- Reinforcement learning algorithms are a set of algorithms and approaches that essentially allow you to give feedback to a model.
強化學習算法是一套算法和方法，從本質上講，它允許你向模型提供反饋。
【TED】Tom Chatfield：遊戲獎勵大腦的7種方式 7 ways games reward the brain
- It's minute reinforcement.
如果你願意可以把它想像為
- It all feeds in as minute reinforcement.
這些都是通過小小的鼓勵實現的。
AI 不斷進步，現在還出現 AGI？！究竟 AI 跟 AGI 有什麼差別，讓專家解釋給你聽！(What's the Difference: AI and AGI?)
- Another type of AI is reinforcement learning, which learns through trial and error, primarily using feedback from its own actions and experiences, rather than from explicit teaching.
另一種人工智能是強化學習，它主要利用自身行動和經驗的反饋，而不是明確的教學，通過嘗試和錯誤來學習。
- In the context of self-driving cars, reinforcement learning is used in simulated environments where the system learns to make decisions by experiencing virtual scenarios, thus improving its algorithms before ever hitting the real road.
在自動駕駛汽車方面，強化學習用於模擬環境，系統通過體驗虛擬場景來學習決策，從而在真正上路之前改進算法。
人工智慧越發先進、各種聊天機器人爭相出現，我們該如何預防 AI 亂象呢？（How to stop AI going rogue）
- The most notable is called "reinforcement learning from human feedback", or RLHFs.
最值得注意的是被稱「人類反饋強化學習」的 RLHF。
- The most notable is called reinforcement learning from human feedback, or RLHFs.
其運作方式是，標籤人員會被要求提供幾個問題給模型，如果輸出不安全，他們會告知模型，
宣傳遊戲：芝麻信用 - 遊戲化的真正危險 - Extra Credits (Propaganda Games: Sesame Credit - The True Danger of Gamification - Extra Credits)
- The world we're stepping into, instead, uses positive reinforcement to promote being subservient to the will of the regime.
像群體的自我審查, 刑期或者暗殺
- Positive reinforcement works wonders.
政府操縱著這種『再教育』, 卻是藉著個人的朋友, 同學, 還有那些想要得高分的關係人們的手
工廠AI的崛起！ (The Rise of AI in Factories)
- Today we use very sophisticated algorithms all the way from deep neural networks through reinforcement learning all the way now to transformers and the more modern generative AI solutions in order to not just tell you something is wrong, but also help guide the technician in exactly what they need to do in order to fix the issues.
如今，我們使用非常複雜的算法，從深度神經網絡到強化學習，一直到變壓器和更現代的生成式人工智能解決方案，不僅能告訴你出了什麼問題，還能幫助指導技術人員如何準確解決問題。
- Today, we use very sophisticated algorithms all the way from deep neural networks through reinforcement learning, all the way now to transformers and the more modern generative AI solutions in order to not just tell you something is wrong but also help guide the technician in exactly what they need to do in order to fix the issues.
事實上，我們的目標不是取代人們，而是用新工具、更現代化的工具武裝他們。
