ç±³é¦é½ã¯ã·ã³ãã³ã§ãé³ã«è³ããã°ã ã¦ã女æ§ï¼2018å¹´5æ16æ¥æ®å½±ãè³æåçï¼ã(c)AFP PHOTO / Eric BARADAT ã5æ17æ¥ AFPããYannyï¼ã¤ãã¼ï¼ãã¨ãLaurelï¼ãã¼ã¬ã«ï¼ããããªãã¯ã©ã¡ãã«èãããï¼ââ2é³ç¯ã®ã¿ã®çãé³å£°ãæ·»ãã¦ããåãããããã¤ãã¿ã¼ï¼Twitterï¼ã®æ稿ããã¤ã³ã¿ã¼ãããä¸ã§ççºçãªé¨ããå·»ãèµ·ããã¦ãããã¦ã¼ã¶ã¼ãã¡ã®åçã¯çã£äºã¤ã«åããã両è ã¨ãè²ãæ°é ã¯ãªãã 16æ¥ä»ã®ç±³ç´ãã¥ã¼ã¨ã¼ã¯ã»ã¿ã¤ã ãºï¼New York Timesï¼ã«ããã¨ããã®é³å£°ã¯ãã¨ãã¨ç±³ã¸ã§ã¼ã¸ã¢å·ã®é«æ ¡çãã¼ã©ã³ãã»ãµãã¼ï¼Roland Szaboï¼ããï¼18ï¼ãå¦æ ¡ã®èª²é¡ã§ãªã³ã©ã¤ã³ã®èªå½ï¼ããï¼è¾å ¸ã使ç¨ä¸ã«è¦ã¤ãã¦é²é³ããã½ã¼ã·ã£ã«ãã¥ã¼ã¹ãµã¤ãã®ã¬ãã£ããï¼Redditï¼ã«æ稿ãããã®ã ç´åãã¡ã«èãããã¨ããããYannyãã¨ãL
ï¼»éå ±ï¼½AIã人éã¨èªç¶ãªä¼è©±ãããé»è©±ã§ç¾å®¹é¢ãã¬ã¹ãã©ã³ãäºç´ãGoogleããã¢ãå ¬éãGoogle I/O 2018 Googleã¯ããµã³ãã©ã³ã·ã¹ã³ã§éå¬ä¸ã®ã¤ãã³ããGoogle I/O 2018ãã«ããã¦ã人éã®ããã«èªç¶ã«ä¼è©±ãããAIã«ãã£ã¦é»è©±ã§ç¾å®¹é¢ãã¬ã¹ãã©ã³ãäºç´ãããã¢ãå ¬éãã¾ããã Google I/Oåæ¥ã®åºèª¿è¬æ¼ã§ãã£ã¨ãä¼å ´ãçãä¸ãã£ãã®ã¯ãGoogleããGoogle Duplexãã¨å¼ã¶AIã®æè¡ãç¨ãã¦ãAIãå®éã«ç¾å®¹é¢ãã¬ã¹ãã©ã³ã«é»è©±ããããäºç´ãè¡ãé³å£°ãå ¬éããå ´é¢ã§ããã ã·ããªãªã¨ãã¦ã¯ãã¦ã¼ã¶ã¼ãGoogle Assistantã«ãæ¥é±ã®ç«ææ¥ã10æãã12æã®ããã ã§ç¾å®¹é¢ãäºç´ãã¦ãã¨ä¾é ¼ã ããã¨Google Assistantã¯ä¾é ¼ãåãã¦Google Duplexã®æ©è½ãç¨ããããã¯ã°ã©ã¦ã³ãã§ç¾å®¹é¢ãäºç´ãã¹ãé»è©±
Figure from our paper: given any waveform, we can modify it slightly to produce another (similar) waveform that transcribes as any different target phrase. We have constructed targeted audio adversarial examples on speech-to-text transcription neural networks: given an arbitrary waveform, we can make a small perturbation that when added to the original waveform causes it to transcribe as any phras
å°å·ãã ã¡ã¼ã«ã§éã ããã¹ã HTML é»åæ¸ç± PDF ãã¦ã³ãã¼ã ããã¹ã é»åæ¸ç± PDF ã¯ãªããããè¨äºãMyãã¼ã¸ããèªããã¨ãã§ãã¾ã 人éã®ããã«é³å£°ãç解ããé³å£°èªè製åã®éçºã§AmazonãAppleãGoogleãIBMãªã©ã¨ç«¶åããMicrosoftããé³å£°èªèã®èª¤ãçã®ä½ãã§ä¸çè¨é²ãéæããã¨çºè¡¨ããã Microsoftã«ããã¨ãMicrosoft Researchã®é³å£°ç§å¦è ã¯æ¥çæ¨æºã®è©ä¾¡ã®ä¸ã§ãåèªèª¤ãçï¼WERï¼6.3ï¼ ãè¨é²ããã¨ãããå©ç¨ããã®ã¯ãå°æ¥çã«ãCortanaããå¼·åããé³å£°èªèæè¡ã ã ããã¾ã§è¨é²ãä¿æãã¦ããã®ã¯IBMã®ãWatsonããã¼ã ã§ãWERã¯6.9ï¼ ã ã£ããIBMã¯2015å¹´ã«éæãã8ï¼ ã¨ããè¨é²ãæ´æ°ããã°ããã ã£ãã Microsoftã¨IBMã¯ãç±³å½æé9æ12æ¥ã¾ã§ãµã³ãã©ã³ã·ã¹ã³ã§éå¬ãããã«ã³ãã¡ã¬
Research WaveNet: A generative model for raw audio Published 8 September 2016 Authors Aäron van den Oord, Sander Dieleman This post presents WaveNet, a deep generative model of raw audio waveforms. We show that WaveNets are able to generate speech which mimics any human voice and which sounds more natural than the best existing Text-to-Speech systems, reducing the gap with human performance by ove
é³å£°å ¥åããªããªãæ®åããªãçç±ã¨ãã¦ãçºè©±ããªããã°ãªããªãã¨è¨ãç¹ããããå¨ãã«èª°ãããªãç°å¢ãªãããã§ãè¯ãããå¨ãã«ä»äººãããå ´åã«ç¬ãè¨ã®ããã«çºè©±ããã®ã¯ãªããªãé£ãããå¨ãã®äººã«è¿·æã«ãªããã¨ãåé¡ã ããå ¥åå 容ãå¨ãã®äººã«ç¥ããã¨ãããã¨ãåé¡ã§ãããããã«é¨é³ä¸ã ã¨ãã ã§ããæªãé³å£°èªèçãããã«ä¸ããã¨è¨ãåé¡ããããé³å£°èªèã®æ大ã®æ¬ ç¹ã¯å£°ãåºããã¨ã«ãã£ãã®ã ã å¥è¯å 端ç§å¦æè¡å¤§å¦é¢å¤§å¦é³æ å ±å¦çå¦è¬åº§ 鹿éç 究室ã§éçºããããéå¯è´ã¤ã¶ããèªèï¼ç¡é³å£°èªèï¼ãã¯ãããã®èª²é¡ã解決ããé³å£°å ¥åã®å©ä¾¿æ§ã大ããåä¸ãããå¯è½æ§ããã¤æè¡ã§ãããéå¯è´ã¤ã¶ããèªèã¯ãåã声ãããå°ããªç¬¬ä¸è ã«è´åä¸è½ãªå£°å¸¯ã®æ¯åãä¼´ããªãã¤ã¶ãã(Non-Audible Murmur:NAM)ã®ä½å ä¼å°é³ããè³ã®è£å´ã«è£ çãã身ä½è¡¨æ¥çåãã¤ã¯ããã©ã³ã«ãã£ã¦èªèãããã®ã ï¼ç¡é³å£°èª
2. ãããã 1 ⢠ãªããããããããªããã©å¥½è©ã ã£ãããã ⢠ã§ã â ãã¿ããããï½ï½ï½ â å¾åãæå³ä¸æ â ã¤ã¼ãããç®grepè¨ããªãã§ãã ⢠ã¨ããããã³ãã⦠⢠é ããããã¨ããã®ããããããã? â å ¨ç¶ãããããªãã! ⢠ãã人ã«ãã£ã¦ã¯ãããã«ãªãè³æã§ãï¼ï¼ãã¨ããã³ ã¡ã³ãããã£ãã 2 / 83
Pythonã§é³å£°ä¿¡å·å¦çï¼2011/05/14ï¼ ä»åã¯ãã¹ãã¯ããã°ã©ã ã使ã£ã¦éãã§ã¿ã¾ããããµã¦ã³ãã¹ãã¯ããã°ã©ã ã£ã¦ä½ã£ã¦äººã¯ãã¡ãã¸ãPythonã§ãµã¦ã³ãã¹ãã¯ããã°ã©ã ï¼2011/10/1ï¼ã ãªãã¨ã¹ãã¯ããã°ã©ã ã«ä»»æã®ç»åãåãè¾¼ãã é³å£°ãä½ããã¨ã®ãã¨ããããªãµãã«ã Aphex Twinã¨ãããã¥ã¼ã¸ã·ã£ã³ã®Equationã¨ããæ²ã«åãè¾¼ã¾ããç»åã ããã§ãããããªé¢¨ã«ç»åãã¹ãã¯ããã°ã©ã ã«ç°¡åã«åãè¾¼ããã¼ã«ãããã¤ãè¦ã¤ããã®ã§éãã§ã¿ã¾ãã Coagula Coagulaã¨ãããã¼ã«ã§ç»åãã¹ãã¯ããã°ã©ã ã«åãè¾¼ãã é³å£°ãä½ãã¾ããCoagulaãèµ·åããããã£ã³ãã¹ãåºã¦ããã®ã§ããã«ä»»æã®çµµãæãã¾ããF5ãã¼ã§ç»åãã¹ãã¯ããã°ã©ã ã«å¤æããFileã¡ãã¥ã¼ã®Save Sound As...ã§WAVEãã¡ã¤ã«ã¨ãã¦é³å£°ãä¿åã§ãã¾ããWAVEã
é³å£°èªèã¯äººå·¥ç¥è½ã®åéã®ä¸ã§ãç¬èªã®é²åãã¨ããåéã§ï¼è¯ããæªããã¬ã©ãã´ã¹çã¨è¨ãããããã¾ãï¼ ç¹ã«å¤§èªå½é£ç¶é³å£°èªèãå®ç¾ããæ¢åã®ã½ããã¦ã¨ã¢ã¯å¤§è¦æ¨¡ãã¤è¤éã§ï¼é³å£°èªèã®å°é家ã§ãããå ¨ä½ãç解ãã¦æ¹è¯ãå ãããã¨ã¯å¿ ããã容æã§ã¯ããã¾ããï¼ãã®ãã¨ã¯è¿é£åéã¨é³å£°èªèã³ãã¥ããã£ãåæããéå£ã¨ããªã£ã¦ãã¾ãï¼ ãããé³å£°èªèãå®ç¾ããã¢ã«ã´ãªãºã èªä½ã¯ï¼åºæ¬çã«ã¯å®ã¯ããã»ã©é£è§£ãªãã®ã§ã¯ããã¾ããï¼ã½ããã¦ã¨ã¢ãè¤éãªã®ã¯ï¼å¤åã«è¨ç®éãã¡ã¢ãªéåæ¸ã®ããã®æ§ã ãªå·¥å¤«ãCã«ä»£è¡¨ãããæç¶ãåããã°ã©ãã³ã°è¨èªã®æ½è±¡åè½åã®éçã«èµ·å ãã¦ãã¾ãï¼ ä»æ¹ï¼ã½ããã¦ã¨ã¢å·¥å¦ã®åéã§ã¯è¤éãªå¦çãã³ã³ãã¯ãã«è¨è¿°å¯è½ãªæ¬¡ä¸ä»£ããã°ã©ãã³ã°ãã©ãã¤ã ã¨ãã¦ï¼ç´ç²é¢æ°åè¨èªãç 究ããã¦ãã¾ãï¼ç´ç²é¢æ°åè¨èªã¯é·ããç 究段éã«çã¾ã£ã¦ãã¾ãããï¼è¿å¹´ã¯Haskellãªã©å®ç¨æ§ã®é«
æµ·å¤ã§å ã«å ¬éããããã®èªèçã®é«ãã§è©±é¡ã¨ãªã£ãiPhoneã¢ããªãDragon Dictation ããã¤ãã«æ¥æ¬èªã«å¯¾å¿ãApp Storeã§ç¡æã§å ¬éããã¦ãã¾ãã iPhoneã«åãã£ã¦ããã¹ã£ãå 容ãããã¹ãã«å¤æãã¦ãããã®ã§ãã¿ããããã«ã§ã®æåå ¥åãè¦æãªã²ã¨ã§ããç´ æ©ããã¤ã¹ãã¬ã¹ç¡ã大éã®ã¡ã¢ãé·æã¡ã¼ã«ãæ¸ããã¨ãã§ãã¾ãã ã¢ããªã®ååèµ·åæã«ã使ç¨è¦ç´ã¸ã®æ¿è«¾ãªã©ãæ±ãããã¾ããã¢ãã¬ã¹å¸³ã®é£çµ¡å ãã¢ãããã¼ãããã¨å人åã®èªèçãé«ããªãããã§ãããå¿ é ã§ã¯ããã¾ããã ãã¡ããéçºå ã«ãããã¢åç»ã ãã¤ã¯ããã®é³å£°å ¥åããµã¼ãã¼ã«éä¿¡ãã¦å¦çãããããWi-FIã¾ãã¯3Gåç·ã§ãããã¯ã¼ã¯ã«æ¥ç¶ãã¦ããå¿ è¦ãããã¾ãã 使ç¨æ³ã¯ããã£ã¦ã·ã³ãã«ã§ãé²é³ãã¿ã³ãæ¼ãã¦ãã¤ã¯ã«åãã£ã¦ããã¹ãã ãã§ãã ãå®äºããã¿ã³ãã¿ããããã¨ãã¼ã¿ããµã¼ãã¼ã«éä¿¡ããã
è¨èªã®ãèµ·æºã¨é²åããæ¢ãç 究ï¼ã人éã¯è¨èªã®å®¿ä¸»ã«ãããªãã 2008å¹´8æ 7æ¥ ãµã¤ã¨ã³ã¹ã»ãã¯ããã¸ã¼ ã³ã¡ã³ãï¼ ãã©ãã¯ãã㯠(0) Brandon Keim Photo: Alpha éºä¼åã®ãããªãµãã¾ããè¦ãããã¼ã ããææ©ä½ã®ããã«é²åããæåã«é¢ããè°è«ã¯ã«ãããã ã[ãã¼ã ã¯åç©è¡åå¦è ãªãã£ã¼ãã»ãã¼ãã³ã¹æ°ãå±ããæ¦å¿µã§ããæåã®è¤è£½éºä¼åããæåå ã®ãæ å ±ããééºä¼çã«æ¿ç¶ãããèªç¶é¸æããããæ§åãããã¼ã¦ã£ã³é²åè«ãåºç¤ã¨ããéºä¼åã®é²åã®éç¨ã«ãªããããã¨ããéºä¼åã«ç¸å½ããä»®æ³ã®ä¸»ä½] ãããä»ã®ã¨ãããçç©å¦çãªãã®ã§ã¯ãªãé²åã«é¢ããå¦åçãªçè«ã¯åå¨ãã¦ããªããã ããããããç¶æ³ããå¤åããå¯è½æ§ãåºã¦ããã 7æ28æ¥(ç±³å½æé)ä»ã®ãç±³å½ç§å¦ã¢ã«ããã¼ç´è¦ã(PNAS)ã«æ²è¼ãããç 究è«æã®ä¸ã§ãè¨èªå¦è ãã¡ã¯ãå®é¨ç°å¢ã§äººå·¥çã«ä½æããã
é³å£°èªèãç´¹ä»ãããã¼ã¸ ã¨ã«ããããã§ã¯ã ã ãã ãã¨ãé³å£°èªèãã¨ããã¢ããç´¹ä»ãã¾ãã å ¨ä½ãï¼ã»ã¼ï¼ãã®ãã¼ã¸ä¸æã«åã¾ã£ã¦ãã¾ãã ã¨ããã©ããã«ãªã³ã¯ãããã¾ããã ãã®ãªã³ã¯å ã«ã¯ã é£ãã話ããã¼ã話ã¿ãããªãã®ãã¡ããã¡ããã¨ããã¾ãã ã¨ããã§ã話ãç°¡åã«ããããã«ã ã¡ãã£ã¨åãæ··ãã¦ãã¾ãã ããã§ãªãã¨ããã®ãããè¤éãªè©±ã«ãªãã®ã§ã é³å£°èªèã£ã¦ãªã«ã ç°¡åã«è¨ã£ã¦ãã¾ãã°ã 人éãåã£ã声ãæ©æ¢°ãæåã«ç´ããã¨ã§ãã å³ã§æãã¨ãããªæãã§ãã å·¦å´ãé³å£°æ³¢å½¢ï¼ã¤ã¾ãã声ãå³ã«è¡¨ãã¦ããï¼ã§ã å³å´ããããã²ãããªã«ç´ãããã®ã§ãã å·¦å´ã®é³å£°æ³¢å½¢ãå°ã詳ããè¦ã¦ã¿ã 人éã¯è³ã§é³ãèãã¾ããã æ©æ¢°ã¯ãã¤ã¯ã§é³ãèãã¾ãã ããã¦ããã¤ã¯ã§åé²ãããé³ããã®ã¾ã¾è¡¨ç¤ºãããã¨ã ä¸ã®ãããªæãã«ãªãã¾ãã 横軸ãæå»ã§ã縦軸ãæ¯å¹ ã§ãã é³å£°ã¨ããã®ã¯ãããã
åæ山大å¦ã¡ãã£ã¢ã¤ã³ããªã¸ã§ã³ã¹ç 究室ï¼è¥¿æç«ä¸ç 究室ï¼â æ¬ç 究室ã§ã¯ãé³å£°ãç»åãåç»ãèªç¶è¨èªãã»ã³ãµãªã©ã®ãã¡ãã£ã¢ããæ´»ç¨ããèªèã»ç解ã®ç¥çæè¡ã«ãã£ã¦ã人ã¨äººã人ã¨æ©æ¢°ã®ååãè±ããªãã®ã«ããããã®æè²ã»ç 究ãè¡ãã¾ãã 2019年度ããæ´»åãã¦ããç 究室ã§ãã æ¬ç 究室ã®å¦é¨çã¯ãã·ã¹ãã å·¥å¦é¨ ã¡ãã£ã¢ãã¶ã¤ã³ã¡ã¸ã£ã¼ã«æå±ãã¦ãã¾ãã大å¦é¢çã®æå±ã¯ãã·ã¹ãã å·¥å¦ç ç©¶ç§ ç¥çã¢ããªã³ã°ã¯ã©ã¹ã¿ã§ãã ãªããæ¬ç 究室ã¯ã西æãæå±ï¼æ å½ï¼ãããã¼ã¿ã»ã¤ã³ããªã¸ã§ã³ã¹æè²ç 究é¨éåã³å¦çèªä¸»åµé æ¯æ´é¨éï¼ã¯ãªã¨ï¼ã¨ãååãããã¨ã§æè²ã»ç 究活åãé²ãã¦ãã¾ãã ä»å¾ã¯ã社ä¼ã¤ã³ãã©ããã£ã¯ã¹å¦ç°ã®å¦çã®åå ¥ãè¡ãäºå®ã§ãã â ãç¥ããâ 2023/12/9 大å¦é¢çï¼M2ï¼ã®æ¥ç°ãããå½ç«ç 究éçºæ³äººæ å ±éä¿¡ç 究æ©æ§ï¼NICTï¼ã®ç 究å¡ã¿ãªãã¾ã¨å ±èã§çºè¡¨ãããè¤æ°è©±è
ãªãªã¼ã¹ãé害æ å ±ãªã©ã®ãµã¼ãã¹ã®ãç¥ãã
ææ°ã®äººæ°ã¨ã³ããªã¼ã®é ä¿¡
å¦çãå®è¡ä¸ã§ã
j次ã®ããã¯ãã¼ã¯
kåã®ããã¯ãã¼ã¯
lãã¨ã§èªã
eã³ã¡ã³ãä¸è¦§ãéã
oãã¼ã¸ãéã
{{#tags}}- {{label}}
{{/tags}}