EBS Morning Special 5월 13일 토요일: extend an olive branch

 

 

 

1. IDIOM PACKAGE

 

- pomp and circumstance 

 impressive formal activities or ceremonies

거창한 의식, 위풍당당

cf) 엘가의 위풍당당 행진곡

 

- put on the dog

to pretend that you are more intelligent, important, or rich than you really are

 

- extend an olive branch 

 an offer of peace and extends back to biblical times, the ancient Romans and Greek mythology

 

Ever wonder what the phrase “extending the olive branch” means?

It is an offer of peace and extends back to biblical times, the ancient Romans and Greek mythology. A dove brought back an olive branch to Noah to signal the flood had ended. In ancient Rome the defeated used to hold up an olive branch for peace.

 

 

 

# 올리브 (출처: laneandco.ca)

 

 

 

 

반응형

 

 

 

 

 

 

2. NEW YORK TIMES

 

AI Is Getting Better at Mind Reading

Oliver Whang

c.2023 The New York Times Company

 

 

Think of the words whirling around in your head: that tasteless joke you wisely kept to yourself at dinner; your unvoiced impression of your best friend’s new partner. Now imagine that someone could listen in.

   

On Monday, scientists from the University of Texas, Austin, made another step in that direction. In a study published in the journal Nature Neuroscience, the researchers described an artificial intelligence that could translate the private thoughts of human subjects by analyzing fMRI scans, which measure the flow of blood to different regions in the brain.

 

[Expression]

whirl around: 빙빙 돌다, 맴돌다

▶keep ~ to oneself : (~를) 비밀로 간직하다, 마음속에 담아두다

▶listen in = 엿듣다 (= eavesdrop)

※overhear: (우연히) 듣게 되다

fMRI: (Functional magnetic resonance imaging) 기능적 자기 공명 영상(기술) (While an MRI scan allows doctors to examine a patient's organs, tissue, or bones, an fMRI looks at the function of the brain)

 

 

 

 

The study centered on three participants, who came to Huth’s lab for 16 hours over several days to listen to “The Moth” and other narrative podcasts. As they listened, an fMRI scanner recorded the blood oxygenation levels in parts of their brains. The researchers then used a large language model to match patterns in the brain activity to the words and phrases that the participants had heard.

 

[Expression]

▶participant = 참가자

- -ant: ‘~하는 사람[것]’을 나타내는 명사 어미 (ex. servant, assistant, inhabitant etc.)

narrative: 이야기(의)

▶blood oxygenation level = 혈중 산소치

- oxygenation: 산소 처리, 산소 추가 (oxygenate: (동) 산소를 동) 공급하다)

large language model: 대규모 언어 모델(LLM) (GPT나 Bard 등이 이에 속함)

※언어 모델(Language Model, LM): 언어를 이루는 구성요소(글자, 형태소, 단어, 문장, 문단 등)에 확률 값을 부여하고 다음 구성 요소를 예측하거나 생성하는 모델 (더 쉽게 표현하면: 통계적으로 가장 자연스러운 단어들의 조합을 찾아내어 문장을 만들어내는 모델)

 

 

In their study, Huth and his colleagues effectively reversed the process, using another AI to translate the participant’s fMRI images into words and phrases. The researchers tested the decoder by having the participants listen to new recordings, then seeing how closely the translation matched the actual transcript.

 

Almost every word was out of place in the decoded script, but the meaning of the passage was regularly preserved. Essentially, the decoders were paraphrasing.

 

Original transcript :

“I got up from the air mattress and pressed my face against the glass of the bedroom window expecting to see eyes staring back at me but instead only finding darkness.”

 

Decoded from brain activity :

“I just continued to walk up to the window and open the glass I stood on my toes and peered out I didn’t see anything and looked up again I saw nothing.”

 

[Expression]

reverse: (방향·순서 따위를) 역으로 하다, 반대가 되게 하다

▶decoder = 해독기, 해독장치 (여기서는 별도로 추가 사용한 AI)

- decode: (암호 등을) 해독하다

▶out of place = 제자리에 있지 않은, 맞지 않는

the meaning of the passage was regularly preserved: 구절의 의미는 규칙적으로 보존되었다 → 해당 구절의 의미는 일정하게 유지되었다

▶paraphrase = (알기 쉽게) 다른 말로 바꾸어 표현하다

 

 

Finally the subjects watched a brief, silent animated movie, again while undergoing an fMRI scan. By analyzing their brain activity, the language model could decode a rough synopsis of what they were viewing — maybe their internal description of what they were viewing.

 

The result suggests that the AI decoder was capturing not just words but also meaning. “Language perception is an externally driven process, while imagination is an active internal process,” Nishimoto said. “And the authors showed that the brain uses common representations across these processes.”

 

[Expression]

▶synopsis = (글·희곡 등의) 개요, (소설· 영화 따위의) 대강의 줄거리

▶the brain uses common representations across these processes = 두뇌는 언어를 쓸 때와 생각을 할 때, 공통된 표현을 사용한다

 

 

This language-decoding method had limitations, Huth and his colleagues noted. For one, fMRI scanners are bulky and expensive. Moreover, training the model is a long, tedious process, and to be effective it must be done on individuals. When the researchers tried to use a decoder trained on one person to read the brain activity of another, it failed, suggesting that every brain has unique ways of representing meaning.

 

[Expression]

▶bulky = 부피가 큰, 덩치가 큰

▶tedious = 지루한, 싫증 나는

 

it must be done on individuals: (이번 방법이 작동하려면) 개인 단위로 적용돼야 한다 (= 이 방법을 불특정 다수에게 일괄 적용할 수는 없다)

 

 

Participants were also able to shield their internal monologues, throwing off the decoder by thinking of other things. AI might be able to read our minds, but for now it will have to read them one at a time, and with our permission.

 

[Expression]

▶shield = (명) 방패, (동) 보호하다, 방어하다

monologue: 독백 (※dialogue: 대화 / trialogue: 3자 회담, 자회담 3인극)

▶throw something/somebody off = (고통스러운·짜증스러운 것 등을) 떨쳐 버리다 (to get rid of something you do not want, such as a quality or a condition)