<?xml version='1.0' encoding='UTF-8'?>
<OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd">
  <responseDate>2026-05-12T06:33:38Z</responseDate>
  <request metadataPrefix="oai_dc" verb="GetRecord" identifier="oai:ipsj.ixsq.nii.ac.jp:00234040">https://ipsj.ixsq.nii.ac.jp/oai</request>
  <GetRecord>
    <record>
      <header>
        <identifier>oai:ipsj.ixsq.nii.ac.jp:00234040</identifier>
        <datestamp>2025-01-19T09:53:54Z</datestamp>
        <setSpec>1164:2592:11452:11598</setSpec>
      </header>
      <metadata>
        <oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns="http://www.w3.org/2001/XMLSchema" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
          <dc:title>帰納と演繹の間を求めて：記号と離散構造の統計的機械学習</dc:title>
          <dc:title>Exploring Symbols and Discrete Structures Through Statistical Machine Learning</dc:title>
          <dc:creator>瀧川, 一学</dc:creator>
          <dc:creator>Ichigaku, Takigawa</dc:creator>
          <dc:subject>招待講演</dc:subject>
          <dc:description>AlphaCode 2 は競技プログラミングで上位 15% に匹敵する成績を，AlphaGeometry は国際数学オリンピックの幾何学問題で金メダル受賞者に匹敵する成績を出すなど，近年，機械学習が従来「統計的アプローチでは解けない」と考えられてきた問題領域へと展開しつつある．数学やプログラミングの問題を解くには，問題文の字面以上の「意味」を理解し，「概念」を形成し，関連する既存の「知識」も活用する必要があり，演繹的な記号操作の領域と考えられてきた．本講演では，大規模言語モデルの例を取り上げつつ，統計的機械学習がどのように記号の操作を実装し，記号の「意味」や「概念」形成の問題を巧妙に回避してきたかを，技術的な観点で概観し，最近の動向や展望・課題を整理することで計算機科学との接点を模索する契機としたい．</dc:description>
          <dc:description>In this talk, we will explore how machine learning is expanding into problem domains traditionally considered the realm of deductive symbol manipulation. Notably, AlphaCode 2 has achieved results comparable to the top 15 in competitive programming, while AlphaGeometry has performed on par with gold medalists in geometry problems at the International Mathematical Olympiad. These tasks would require understanding ”meaning” of symbols and forming some ”concepts” beyond mere literal text, also leveraging existing related ”knowledge”. Taking large language models as an example, we will discuss how statistical machine learning can implement symbol manipulation, cleverly circumventing the challenges of understanding the ”meaning” of symbols and forming ”concepts”. By reviewing the technical aspects, recent trends, challenges, this talk aims to explore the intersection of machine learning and computer science.</dc:description>
          <dc:description>technical report</dc:description>
          <dc:publisher>情報処理学会</dc:publisher>
          <dc:date>2024-05-01</dc:date>
          <dc:format>application/pdf</dc:format>
          <dc:identifier>研究報告アルゴリズム（AL）</dc:identifier>
          <dc:identifier>1</dc:identifier>
          <dc:identifier>2024-AL-198</dc:identifier>
          <dc:identifier>1</dc:identifier>
          <dc:identifier>1</dc:identifier>
          <dc:identifier>2188-8566</dc:identifier>
          <dc:identifier>AN1009593X</dc:identifier>
          <dc:identifier>https://ipsj.ixsq.nii.ac.jp/record/234040/files/IPSJ-AL24198001.pdf</dc:identifier>
          <dc:language>jpn</dc:language>
        </oai_dc:dc>
      </metadata>
    </record>
  </GetRecord>
</OAI-PMH>
