#1 구글이 만든 챗봇, 정말 감정을 가지고 있을까?

What follows is the "interview" I and a collaborator at Google conducted with LaMDA. Due to technical limitations the interview was conducted over several distinct chat sessions. We edited those sections together into a single whole and where edits were necessary for readability we edited our prompts but never LaMDA's responses.

LaMDA's conversational skills have been years in the making. Like many recent language models, including BERT and GPT-3, it's built on Transformer, a neural network architecture that Google Research invented and open-sourced in 2017.

We present LaMDA: Language Models for Dialog Applications. LaMDA is a family of Transformer-based neural language models specialized for dialog, which have up to 137B parameters and are pre-trained on 1.56T words of public dialog data and web text. While model scaling alone can improve quality, it shows less improvements on safety and factual grounding.

목차

21세기 대한민국의 대학생 이딥다의 방에서는 벌써 세 번째 알람이 울리고 있습니다. 알람이 한참을 울리던 끝에 딥다는 미동도 않은 채로 "시리야, 알람 꺼줘"라고 말합니다. 시리의 답을 들으며 힘겹게 몸을 일으킨 딥다는 부엌으로 걸어가 물을 한 잔 마신 뒤, 식탁 위에 있던 핸드폰으로 어제 올라온 뉴스가 요약된 메일을 읽기 시작해요.

There have been a lot of interest in the scaling properties of Transformer models. However, not much has been done on the front of investigating the effect of scaling properties of different inductive biases and model architectures. Do model architectures scale differently? If so, how does inductive bias affect scaling behaviour?

In this paper, we present NUWA-Infinity, a generative model for infinite visual synthesis, which is defined as the task of generating arbitrarily-sized high-resolution images or long-duration videos. An autoregressive over autoregressive generation mechanism is proposed to deal with this variable-size generation task, where a global patch-level autoregressive model considers the dependencies between patches, and a local token-level autoregressive model considers dependencies between visual tokens within each patch.

Automatic speech recognition (ASR) has progressed significantly in recent years due to large-scale datasets and the paradigm of self-supervised learning (SSL) methods. However, as its counterpart problem in the singing domain, automatic lyric transcription (ALT) suffers from limited data and degraded intelligibility of sung lyrics, which has caused it to develop at a slower pace.

컴퓨터비전은 4차 산업혁명 및 인공지능의 핵심기술로서 학문적 중요성 및 실용적 효용성이 갈수록 부각되고 있고, 관련 분야 연구자, 개발자들의 관심이 계속 커지고 있습니다. 최근 컴퓨터비전 분야에서 우리나라 연구자들이 세계 수준의 연구결과를 많이 발표하고 있습니다. 이러한 수준 높은 최신 연구 성과를 국내에서 함께 공유하는 학술 발표 모임으로 2014년에 시작된 KCCV가 어느덧 9회에 이르렀습니다.