Have you seen Barack Obama call Donald Trump a “complete dipshit,” or Mark Zuckerberg brag about having “total control of billions of people’s stolen data”? Answer yes and you’ve seen a deepfake. The 21st century’s answer to photoshopping, deepfakes use a form of artificial intelligence (AI) called deep learning to create images of fake events, hence the name deepfake.
The AI firm Deeptrace found 15,000 deepfake videos online in September 2019, a near doubling over nine months. A staggering 96 percent were pornographic, and 99 percent of those mapped faces from female celebrities onto porn stars.
Is it just about videos? No. Deepfake technology can create convincing but entirely fictional photos from scratch.
Photo: AFP 照片：法新社
Audio can be deepfaked too. In March 2019, the chief of a UK subsidiary of a German energy firm paid nearly ￡200,000 into a Hungarian bank account after being phoned by a fraudster who mimicked the German CEO’s voice. Similar scams have reportedly used recorded WhatsApp voice messages.
How do you spot a deepfake? Poor-quality deepfakes are easier to spot. The lip synching might be bad, or the skin tone patchy. However, spotting a deepfake gets harder as the technology improves. In 2018, US researchers discovered that deepfake faces do not blink normally. No surprise there: the majority of images show people with their eyes open, so the algorithms never really learned about blinking. At first it seemed like a silver bullet for the detection problem, but no sooner had the research been published than deepfakes appeared with blinking. Such is the nature of the game: as soon as a weakness is revealed, it is fixed.
(The Guardian)
你看過歐巴馬稱川普為「徹頭徹尾的蠢貨」，或是祖克柏吹噓「完全掌控了數十億人被盜的資料」嗎？若你回答是，那麼你已見識過「deepfake」（深偽）。deepfake是二十一世紀對Photoshop影像處理的回應，使用一種稱為「深度學習」的人工智慧來製作影像以偽造事情，因此而得名。
人工智慧公司Deeptrace二○一九年九月在網路上發現了一萬五千個deepfake影片，在九個月內幾乎翻了一倍。其中高達百分之九十六的比例是色情內容，所顯示的臉孔，百分之九十九是女性，從名人到色情片明星都有。
deepfake只限於影片嗎？不是的。此技術可做出令人信服，卻是完全虛構的照片。
聲音也可以被深度偽造。二○一九年三月，一家德國能源公司的英國子公司負責人接到一通詐騙電話，騙子的聲音模仿德國的CEO，該負責人便將近二十萬英鎊匯入一個匈牙利的銀行帳戶。據報導，類似的詐騙手法也用於錄製WhatsApp語音訊息。
要如何辨識deepfake呢？品質差的deepfake較容易被識破──嘴唇可能不太同步，或者膚色不均勻。但隨著技術的進步，辨識deepfake變得越來越難。二○一八年，美國有研究人員發現deepfake的人臉無法正常地眨眼。這不令人意外：大多數影像所顯示的人臉，眼睛都是睜開的，因此演算法從未真正學習到眨眼。起先眨眼似乎是識破深偽的訣竅。但這項研究發表後不久，會眨眼的deepfake就出現了。這就是遊戲的本質：弱點一旦被發現，就會被修復。
（台北時報林俐凱編譯）
Street lights are often taken for granted until a power outage plunges the world into darkness. When that happens, the value of these lighting installations becomes evident as the world turns into a more dangerous place for pedestrians and motorists alike. The Chinese could claim to be the first to have constructed a crude type of street light. Around 500 BC, residents of Beijing employed a type of street lamp that used hollow bamboo pipes and natural gas vents to create burning torches. Later, ancient Romans adopted lamps fueled by vegetable oil, which relied on slaves to light and
A: Who else is on Billboard’s list: “The 25 greatest pop stars of the 21st Century?” B: No. 15 to 6 are: Miley Cyrus, Justin Timberlake, Nicki Minaj, Eminem, Usher, Adele, Ariana Grande, Justin Bieber, Kanye West and Britney Spears. A: I can’t believe that Adele’s only at No. 10. B: No. 5 to 1 are: Lady Gaga, Drake, Rihanna, Taylor Swift and Beyonce. A: Well, they surely deserve the honor. A: 《告示牌》雜誌的「21世紀最偉大的25位流行歌手」，還有誰上榜啊？ B: 第15至6名是：麥莉希拉、大賈斯汀、妮姬米娜、阿姆、亞瑟小子、愛黛兒、亞莉安娜、小賈斯汀、肯伊威斯特、小甜甜布蘭妮。 A: 真不敢相信愛黛兒只排第10名。 B: 第5至1名是：女神卡卡、德瑞克、蕾哈娜、泰勒絲、碧昂絲。 A: 這幾位真是實至名歸！ （By Eddy Chang, Taipei Times／台北時報張聖恩）
Undersea cables are conductors wrapped in insulating materials and laid on the seabed. Their main functions are telecommunications or power transmission. The core of the undersea cables used for Internet signals is optical fiber, using light to transmit Internet signals. Taiwan’s communications are currently handled by 10 domestic undersea cables and 14 international undersea cables. About 99 percent of Taiwan’s Internet bandwidth relies on undersea cables, making them Taiwan’s “digital lifeline.” The demands on the cables’ bandwidth are only set to increase with the development of artificial intelligence (AI), which relies on the data fed into it. Today, data is
Spoiler alert and shift blame 破梗&甩鍋 在新冠疫情期間，無論是因為封城 (lockdown) 還是居家隔離 (self-isolation at home)，人們關在家中使用網路的時間大增。這也讓一些原本只存在於網路論壇的用語廣為普及。我們來談一下破梗 (spoiler alert) 與甩鍋 (shift blame) 這兩個用語。 有位古典文學教授 Joel Christensen 針對領導統御與疫情控制寫了一篇以古喻今、相當深入的文章：“Plagues follow bad leadership in ancient Greek tales”，文中出現一些講法，可用來翻譯上述的流行語： In the 5th century B.C., the playwright Sophocles begins Oedipus Tyrannos with the title character struggling to identify the cause of a plague striking his city, Thebes. (Spoiler alert: It’s his own bad leadership.) (Joel Christensen, “Plagues follow bad leadership in ancient Greek tales,” The Conversation, March 12, 2020) 作者提到 Oedipus（伊底帕斯）想找出瘟疫何以降臨他的城邦的緣由，加了一句：Spoiler alert: It’s his own bad leadership.（破梗：領導無方）。Spoiler alert 就是「破梗」，如果用在有人洩漏電影劇情的情境中，也可以翻作「小心爆雷」或「劇透警告」。疫情之下，在家看影集、電影成了很多人的娛樂，但要小心劇透 (spoilers)，很多 YouTube 上的影評在開頭也都會說 Spoiler alert!，警告還沒看過電影的觀眾小心爆雷、劇透。 至於「甩鍋」，源自大陸網民用語，通常意指某人犯了錯之後想推卸責任、轉移焦點、甚至讓別人背黑鍋的做法。疫情爆發後，相關網路資訊量爆增，許多中國網民也想找人為這場疫情負責，紛紛呼籲地方政府首長、地方黨書記不要「甩鍋」。 其實，在古代文學《奧德賽》中，就有「將自己的責任怪罪眾神」的說法，試用時下流行的「甩鍋」來重新翻譯： Humans are always blaming the gods for their suffering, but they experience
RSS