Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
Neural Machine Translation (NMT) is an end-to-end learning approach for automated translation, with the potential to overcome many of the weaknesses of conventional phrase-based translation systems. Unfortunately, NMT systems are known to be computationally expensive both in training and in translation inference. Also, most NMT systems have difficulty with rare words. These issues have hindered NM
人気ã®æ³¢ã«ã¯é€†ã‚‰ãˆãšã€ä»Šæ›´ãªãŒã‚‰Pythonを始ã‚ã¦ã¿ãŸ
Pythonã¨ã„ã†ã¨ã€æ—¥æœ¬ã§ã¯é•·ã‚‰ãややマイナーãªå˜åœ¨ã§ã—ãŸãŒã€ã“ã“ã«æ¥ã¦ã„ããªã‚Šã€æœ€ã‚‚é‡è¦ãªãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°è¨€èªžã«ãªã£ãŸæ„ŸãŒã‚ã‚Šã¾ã™ã€‚ã‚‚ã—ã€â€œPRG48â€ã¨ã„ã†ã‚°ãƒ«ãƒ¼ãƒ—ãŒã‚ã£ã¦ã€â€œãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°è¨€èªžç·é¸æŒ™â€ã‚’ã‚„ã£ãŸã¨ã—ãŸã‚‰ã€äººæ°—急上昇ã§1ä½ã«ãªã‚Šãã†ãªå‹¢ã„…。ã™ã¿ã¾ã›ã‚“ã€ã¤ã¾ã‚‰ãªã„例ãˆã‚’ã—ã¦ã—ã¾ã„ã¾ã—ãŸã€‚ Python人気ã®æœ€å¤§ã®ç†ç”±ã¯æœ€è¿‘ã®äººå·¥çŸ¥èƒ½ãƒ–ームã§ã—ょã†ã€‚「TensorFlowã€ã‚„「scikit-learnã€ã¨ã„ã£ãŸæ©Ÿæ¢°å¦ç¿’ã®æœ‰åライブラリãŒPythonå‘ã‘ã«æä¾›ã•ã‚Œã¦ã„ã‚‹ã®ã§ã€ã€Œäººå·¥çŸ¥èƒ½ã‚„機械å¦ç¿’ã‚’å¦ã¶ãªã‚‰Pythonã€ã¨ã„ã†èªè˜ãŒåºƒã¾ã£ã¦ã„ã¾ã™ã€‚ 書店ã®æŠ€è¡“æ›¸å£²ã‚Šå ´ã¸è¡Œãã¨æ©Ÿæ¢°å¦ç¿’ã‚’Pythonã®ã‚³ãƒ¼ãƒ‰ã¨ã¨ã‚‚ã«è§£èª¬ã™ã‚‹æœ¬ãŒä½•å†Šã‚‚並んã§ã„ã¾ã™ã€‚ã“ã®å‡ºç‰ˆä¸æ³ã®ã”時世ã«ã€ãŠä¸–辞ã«ã‚‚ç°¡å˜ã¨ã¯è¨€ãˆãªã„機械å¦ç¿’ã®æœ¬ãŒä½•å†Šã‚‚出ã¦ã„る状æ³ã¯ã€æ—¥æœ¬ã®ã‚½ãƒ•ãƒˆã‚¦ã‚¨ã‚¢æ¥ç•Œã®æœªæ¥ã¯æ¡ˆå¤–明るã„ã€ã¨æ€ã‚ã›
機械å¦ç¿’ ã¯ã˜ã‚よㆠ記事一覧 | gihyo.jp
é‹å–¶å…ƒã®ãƒã‚´ Copyright © 2007-2024 All Rights Reserved by Gijutsu-Hyoron Co., Ltd. ページ内容ã®å…¨éƒ¨ã‚ã‚‹ã„ã¯ä¸€éƒ¨ã‚’ç„¡æ–ã§åˆ©ç”¨ã™ã‚‹ã“ã¨ã‚’ç¦æ¢ã—ã¾ã™â 。個別ã«ãƒ©ã‚¤ã‚»ãƒ³ã‚¹ãŒè¨å®šã•ã‚Œã¦ã„る記事ç‰ã¯ãã®ãƒ©ã‚¤ã‚»ãƒ³ã‚¹ã«å¾“ã„ã¾ã™ã€‚
第2回　確率ã®åˆæ© | gihyo.jp
今回ã¯ã€æ©Ÿæ¢°å¦ç¿’ã§ä½¿ã†ã€Œç¢ºçŽ‡ã€ã®ãŠè©±ã§ã™ã€‚ 確率ã¯ã€çµ±è¨ˆçš„ãªæ©Ÿæ¢°å¦ç¿’ã®ã‚‚ã£ã¨ã‚‚é‡è¦ãªåŸºç¤ŽçŸ¥è˜ã§ã™ã€‚ã¨ã¯ã„ãˆã€ç¢ºçŽ‡ã«ã¤ã„ã¦ã‚¼ãƒã‹ã‚‰èª¬æ˜Žã™ã‚‹ã¨ã„ã†ã®ã¯ç´™æ•°çš„ã«ã‚‚厳ã—ã„ãŸã‚ã€é«˜æ ¡ã®ç¢ºçŽ‡ã‚’å°‘ã—憶ãˆã¦ã„ã‚‹ãらã„（期待値や標準åå·®ãªã©ï¼‰ã‚’å‰æã¨ã—ã€ã€Œâ é«˜æ ¡ã®ç¢ºçŽ‡ã€ã¨ã€Œæ©Ÿæ¢°å¦ç¿’ã®ç¢ºçŽ‡ã€ã®æœ¬è³ªçš„ãªç›¸é•ç‚¹ã«ã¤ã„ã¦ã€å°‘ã—ä¸å¯§ã«è¦‹ã¦ã„ãã€ã¨ã„ã†å½¢ã§é€²ã‚ã¦ã„ãã¾ã™ã€‚ 機械å¦ç¿’ã¨ç¢ºçŽ‡ 最åˆã«ã€æ©Ÿæ¢°å¦ç¿’ã«ã¨ã£ã¦ç¢ºçŽ‡ã¯ã©ã†ã„ã†å½¹å‰²ãªã®ã‹ã‚’確èªã—ã¦ãŠãã¾ã—ょã†ã€‚ 実ã®ã¨ã“ã‚ã€æ©Ÿæ¢°å¦ç¿’ã«ç¢ºçŽ‡ãŒå¿…é ˆã¨ã„ã†ã‚ã‘ã§ã¯ã‚ã‚Šã¾ã›ã‚“。ニューラルãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚„サãƒãƒ¼ãƒˆãƒ™ã‚¯ã‚¿ãƒ¼ãƒžã‚·ãƒ³ãªã©ã®æœ‰åãªæ‰‹æ³•ã‚‚「確率を用ã„ãªã„機械å¦ç¿’ã€ã§ã™ã—ã€ãã®ä»–ã«ã‚‚数多ãã®æ‰‹æ³•ãŒã‚ã‚Šã¾ã™ã€‚ã—ã‹ã—ã€ã€Œâ 確率を用ã„ãªã„機械å¦ç¿’ã€ã®å¤šãã¯ã€ã€Œâ çµæžœã®ãƒ©ãƒ³ã‚ングを作りã¥ã‚‰ã„（評価値ã®å¤§å°ã«æ„味ãŒãªã„â ）â ã€â 「â æ¡ä»¶ãŒç•°ãªã‚‹å ´åˆã®çµæžœã‚’比較ã§ããªã„ã€ãªã©ã®æ¬ 点ãŒã‚ã‚Šã¾
1
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
{{#tags}}- {{label}}
{{/tags}}