VSGenerator/ja.html at main · Bgolearn/VSGenerator · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
<!doctype html>
<html lang="en">
<head>
  <meta charset="utf-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <meta name="description" content="Spatial-adaptive active learning for discovering durable acidic OER catalysts.">
  <title>Spatial-Adaptive Active Learning | Acidic OER</title>
  <style>
    :root { --ink:#10243a; --muted:#64748b; --line:#dce5ec; --paper:#f8fafb; --blue:#1169a9; --cyan:#dff4f5; --orange:#f59e0b; }
    * { box-sizing:border-box; }
    html { scroll-behavior:smooth; }
    body { margin:0; color:var(--ink); background:var(--paper); font-family:Inter, ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", "Noto Sans", sans-serif; line-height:1.55; }
    .wrap { width:min(1120px, calc(100% - 40px)); margin:auto; }
    .topbar { position:absolute; inset:0 0 auto; z-index:2; padding:22px 0; }
    .topbar .wrap { display:flex; justify-content:space-between; align-items:center; gap:16px; }
    .brand { color:white; text-decoration:none; font-weight:750; letter-spacing:.04em; font-size:.82rem; }
    .languages { display:flex; background:rgba(255,255,255,.13); padding:4px; border:1px solid rgba(255,255,255,.25); border-radius:999px; backdrop-filter:blur(8px); }
    .languages button { border:0; color:#e7f4fb; background:transparent; cursor:pointer; padding:5px 10px; border-radius:999px; font:inherit; font-size:.75rem; }
    .languages button.active { color:var(--ink); background:white; font-weight:700; }
    .hero { min-height:610px; display:grid; align-items:end; overflow:hidden; position:relative; color:white; background:linear-gradient(135deg,#062139 0%,#0a486a 55%,#117c9a 100%); }
    .hero::after { content:""; position:absolute; width:680px; height:680px; right:-190px; top:-250px; border:1px solid rgba(255,255,255,.2); border-radius:50%; box-shadow:0 0 0 72px rgba(255,255,255,.045),0 0 0 144px rgba(255,255,255,.035); }
    .hero-content { z-index:1; padding:145px 0 70px; max-width:830px; }
    .eyebrow { display:inline-flex; gap:8px; align-items:center; color:#bceaf0; text-transform:uppercase; letter-spacing:.14em; font-weight:700; font-size:.72rem; }
    .eyebrow::before { content:""; display:block; width:30px; height:1px; background:#62d3da; }
    h1 { font-size:clamp(2.5rem,6vw,5.25rem); line-height:.98; letter-spacing:-.065em; margin:18px 0 23px; max-width:900px; }
    .lead { max-width:660px; font-size:clamp(1rem,1.7vw,1.25rem); color:#e2f2f5; margin:0; }
    .metrics { display:flex; flex-wrap:wrap; gap:1px; margin-top:45px; width:max-content; max-width:100%; background:rgba(255,255,255,.2); }
    .metric { min-width:156px; padding:15px 21px; background:rgba(5,31,52,.58); backdrop-filter:blur(8px); }
    .metric strong { display:block; font-size:1.55rem; letter-spacing:-.04em; color:#f5cf6c; line-height:1.1; }
    .metric span { display:block; margin-top:4px; color:#d5e8ed; font-size:.78rem; }
    main { padding:88px 0 104px; }
    .intro { display:grid; grid-template-columns:.8fr 1.2fr; gap:60px; align-items:start; margin-bottom:98px; }
    .section-tag { color:var(--blue); text-transform:uppercase; letter-spacing:.13em; font-size:.73rem; font-weight:800; }
    h2 { margin:10px 0 0; font-size:clamp(2rem,3.8vw,3.45rem); line-height:1.05; letter-spacing:-.055em; }
    .intro p { margin:0; font-size:1.08rem; color:#3c5063; }
    .intro p + p { margin-top:18px; }
    .highlights { display:grid; grid-template-columns:repeat(3,1fr); gap:18px; margin-bottom:100px; }
    .highlight { background:white; border:1px solid var(--line); padding:27px; border-radius:4px; min-height:200px; box-shadow:0 13px 30px rgba(20,45,67,.04); }
    .number { color:var(--orange); font-size:2.15rem; line-height:1; font-weight:800; letter-spacing:-.06em; }
    .highlight h3 { margin:25px 0 8px; font-size:1rem; }
    .highlight p { margin:0; color:var(--muted); font-size:.91rem; }
    .gallery-head { display:flex; justify-content:space-between; align-items:end; gap:25px; margin-bottom:28px; }
    .gallery-head p { max-width:390px; margin:0; color:var(--muted); font-size:.92rem; }
    .figures { display:grid; grid-template-columns:repeat(12,1fr); gap:20px; }
    figure { overflow:hidden; margin:0; background:white; border:1px solid var(--line); border-radius:5px; }
    figure.workflow { grid-column:span 12; }
    figure.cvae,figure.performance { grid-column:span 6; }
    figure.characterization { grid-column:span 12; }
    figure img { width:100%; display:block; aspect-ratio:16/9; object-fit:cover; object-position:center; transition:transform .45s ease; }
    figure.workflow img { aspect-ratio:1.34/1; object-fit:contain; background:white; }
    figure.characterization img { aspect-ratio:1.55/1; object-fit:cover; object-position:40% center; }
    figure:hover img { transform:scale(1.015); }
    figcaption { padding:15px 17px 17px; color:#4b6173; font-size:.86rem; }
    figcaption b { display:block; color:var(--ink); margin-bottom:2px; }
    .citation { margin-top:74px; padding:27px 0 0; border-top:1px solid var(--line); display:flex; justify-content:space-between; align-items:flex-start; gap:24px; color:var(--muted); font-size:.85rem; }
    .citation p { margin:0; max-width:760px; }
    .paper-link { flex:none; display:inline-block; color:var(--blue); border:1px solid #a9cde2; padding:9px 13px; border-radius:3px; text-decoration:none; font-weight:700; white-space:nowrap; }
    .paper-link:hover { color:white; background:var(--blue); }
    @media (max-width:720px) { .wrap { width:min(100% - 28px,1120px); } .topbar { padding:17px 0; } .brand { font-size:.7rem; } .hero { min-height:650px; } .hero-content { padding:135px 0 45px; } .metric { min-width:calc(50% - 1px); padding:13px 16px; } main { padding:58px 0 70px; } .intro { grid-template-columns:1fr; gap:25px; margin-bottom:54px; } .highlights { grid-template-columns:1fr; margin-bottom:58px; } .gallery-head,.citation { display:block; } .gallery-head p { margin-top:15px; } figure.cvae,figure.performance,figure.characterization { grid-column:span 12; } .citation { margin-top:48px; } .paper-link { margin-top:16px; } }
  </style>
</head>
<body>
  <header class="hero">
    <nav class="topbar"><div class="wrap"><a class="brand" href="#top">SPATIAL-ADAPTIVE AL</a><div class="languages" aria-label="Language selector"><button class="active" data-lang="en">EN</button><button data-lang="zh">中</button><button data-lang="ja">日</button><button data-lang="ko">한</button></div></div></nav>
    <div class="wrap hero-content" id="top">
      <span class="eyebrow" data-i18n="eyebrow">MATERIALS DISCOVERY</span>
      <h1 data-i18n="title">Find the catalyst.<br>Cut the wait.</h1>
      <p class="lead" data-i18n="lead">A spatial-adaptive active learning strategy identifies a Cu–RuO₂ catalyst that couples exceptional acidic OER activity with long-term durability.</p>
      <div class="metrics">
        <div class="metric"><strong>177 mV</strong><span data-i18n="metric1">Overpotential @ 10 mA cm⁻²</span></div>
        <div class="metric"><strong>625 h</strong><span data-i18n="metric2">Stable operation in acid</span></div>
        <div class="metric"><strong>78×</strong><span data-i18n="metric3">Stability improvement</span></div>
      </div>
    </div>
  </header>
  <main class="wrap">
    <section class="intro">
      <div><span class="section-tag" data-i18n="introTag">THE CHALLENGE</span><h2 data-i18n="introTitle">Optimize what is fast.<br>Then optimize what lasts.</h2></div>
      <div><p data-i18n="intro1">Activity can be measured in minutes; durability testing can take hundreds of hours. Treating both objectives equally slows discovery.</p><p data-i18n="intro2">This work links two targeted learning stages: it first finds low-overpotential recipes, then constructs a focused candidate space to search for stability.</p></div>
    </section>
    <section class="highlights" aria-label="Highlights">
      <article class="highlight"><div class="number">19,200 → 785</div><h3 data-i18n="highlight1Title">A search space with focus</h3><p data-i18n="highlight1Text">A conditional variational autoencoder concentrates the original recipe space into promising low-overpotential candidates.</p></article>
      <article class="highlight"><div class="number">23 / 11</div><h3 data-i18n="highlight2Title">Experiments / stability tests</h3><p data-i18n="highlight2Text">The closed loop reaches the design target with only 11 long-duration durability measurements.</p></article>
      <article class="highlight"><div class="number">4,734 h</div><h3 data-i18n="highlight3Title">Testing time saved</h3><p data-i18n="highlight3Text">At least 12 stability tests—and approximately 4,734 testing hours—are avoided during discovery.</p></article>
    </section>
    <section>
      <div class="gallery-head"><div><span class="section-tag" data-i18n="resultsTag">RESULTS AT A GLANCE</span><h2 data-i18n="resultsTitle">A smarter route to durable performance.</h2></div><p data-i18n="resultsText">The workflow narrows the question, the learning loop finds the answer, and characterization confirms the resulting Cu–RuO₂ catalyst.</p></div>
      <div class="figures">
        <figure class="workflow"><img src="docs/spatial-adaptive-workflow.png" alt="Diagram of the spatial-adaptive active learning workflow"><figcaption><b data-i18n="workflowTitle">Two-stage spatial-adaptive learning</b><span data-i18n="workflowText">Bayesian optimization for activity is followed by CVAE-guided stability optimization.</span></figcaption></figure>
        <figure class="cvae"><img src="docs/active-learning-cvae.png" alt="Active learning and conditional variational autoencoder results"><figcaption><b data-i18n="cvaeTitle">Learning the promising region</b><span data-i18n="cvaeText">Iterative experiments improve activity predictions before the CVAE refines the stability search space.</span></figcaption></figure>
        <figure class="performance"><img src="docs/catalyst-performance.png" alt="Catalyst stability and electrochemical performance results"><figcaption><b data-i18n="performanceTitle">Performance validated</b><span data-i18n="performanceText">Cu–RuO₂ reaches 177 mV overpotential and remains stable for 625 hours in acidic conditions.</span></figcaption></figure>
        <figure class="characterization"><img src="docs/catalyst-characterization.png" alt="Microscopy, spectroscopy, and explainability results"><figcaption><b data-i18n="characterizationTitle">Structure and insight</b><span data-i18n="characterizationText">Microscopy, diffraction, spectroscopy, and model interpretation connect discovery to material evidence.</span></figcaption></figure>
      </div>
    </section>
    <footer class="citation"><p data-i18n="citation">B. Cao et al. “Spatial-adaptive active learning identifies ultra-durable and highly active catalysts for acidic oxygen evolution reaction.” <i>Science Bulletin</i> 71 (2026), 1084–1092.</p><a class="paper-link" href="docs/main.pdf" target="_blank" rel="noopener" data-i18n="paper">Read the paper ↗</a></footer>
  </main>
  <script>
    const copy = {
      en: { eyebrow:'MATERIALS DISCOVERY', title:'Find the catalyst.<br>Cut the wait.', lead:'A spatial-adaptive active learning strategy identifies a Cu–RuO₂ catalyst that couples exceptional acidic OER activity with long-term durability.', metric1:'Overpotential @ 10 mA cm⁻²', metric2:'Stable operation in acid', metric3:'Stability improvement', introTag:'THE CHALLENGE', introTitle:'Optimize what is fast.<br>Then optimize what lasts.', intro1:'Activity can be measured in minutes; durability testing can take hundreds of hours. Treating both objectives equally slows discovery.', intro2:'This work links two targeted learning stages: it first finds low-overpotential recipes, then constructs a focused candidate space to search for stability.', highlight1Title:'A search space with focus', highlight1Text:'A conditional variational autoencoder concentrates the original recipe space into promising low-overpotential candidates.', highlight2Title:'Experiments / stability tests', highlight2Text:'The closed loop reaches the design target with only 11 long-duration durability measurements.', highlight3Title:'Testing time saved', highlight3Text:'At least 12 stability tests—and approximately 4,734 testing hours—are avoided during discovery.', resultsTag:'RESULTS AT A GLANCE', resultsTitle:'A smarter route to durable performance.', resultsText:'The workflow narrows the question, the learning loop finds the answer, and characterization confirms the resulting Cu–RuO₂ catalyst.', workflowTitle:'Two-stage spatial-adaptive learning', workflowText:'Bayesian optimization for activity is followed by CVAE-guided stability optimization.', cvaeTitle:'Learning the promising region', cvaeText:'Iterative experiments improve activity predictions before the CVAE refines the stability search space.', performanceTitle:'Performance validated', performanceText:'Cu–RuO₂ reaches 177 mV overpotential and remains stable for 625 hours in acidic conditions.', characterizationTitle:'Structure and insight', characterizationText:'Microscopy, diffraction, spectroscopy, and model interpretation connect discovery to material evidence.', citation:'B. Cao et al. “Spatial-adaptive active learning identifies ultra-durable and highly active catalysts for acidic oxygen evolution reaction.” <i>Science Bulletin</i> 71 (2026), 1084–1092.', paper:'Read the paper ↗' },
      zh: { eyebrow:'材料发现', title:'找到催化剂。<br>缩短等待。', lead:'空间自适应主动学习策略发现了兼具优异酸性 OER 活性与长期耐久性的 Cu–RuO₂ 催化剂。', metric1:'10 mA cm⁻² 下的过电位', metric2:'酸性条件下稳定运行', metric3:'稳定性提升', introTag:'研究挑战', introTitle:'先优化快速指标。<br>再优化持久性能。', intro1:'活性可在数分钟内测得，而耐久性测试往往需要数百小时。将二者等权处理会拖慢发现进程。', intro2:'本研究连接两个目标明确的学习阶段：先寻找低过电位配方，再构建聚焦的候选空间以搜索稳定性。', highlight1Title:'聚焦的搜索空间', highlight1Text:'条件变分自编码器将原始配方空间聚焦为具有低过电位潜力的候选集。', highlight2Title:'实验数 / 稳定性测试数', highlight2Text:'闭环实验仅用 11 次长时耐久性测试便达成设计目标。', highlight3Title:'节省测试时间', highlight3Text:'发现过程中至少避免 12 次稳定性测试，节省约 4,734 小时测试时间。', resultsTag:'核心结果', resultsTitle:'通向持久性能的更聪明路径。', resultsText:'流程缩小问题范围，学习闭环找到答案，表征验证了最终的 Cu–RuO₂ 催化剂。', workflowTitle:'两阶段空间自适应学习', workflowText:'先以贝叶斯优化提升活性，再以 CVAE 引导优化稳定性。', cvaeTitle:'学习有潜力的区域', cvaeText:'迭代实验先提升活性预测能力，再由 CVAE 细化稳定性搜索空间。', performanceTitle:'性能得到验证', performanceText:'Cu–RuO₂ 的过电位为 177 mV，并在酸性条件下稳定运行 625 小时。', characterizationTitle:'结构与洞见', characterizationText:'显微、衍射、光谱与模型解释将发现结果与材料证据相连。', citation:'B. Cao 等，《空间自适应主动学习发现超耐久、高活性酸性析氧反应催化剂》。<i>Science Bulletin</i> 71 (2026), 1084–1092。', paper:'阅读论文 ↗' },
      ja: { eyebrow:'材料探索', title:'触媒を見つける。<br>待ち時間を減らす。', lead:'空間適応型アクティブラーニングにより、優れた酸性 OER 活性と長期耐久性を併せ持つ Cu–RuO₂ 触媒を発見しました。', metric1:'10 mA cm⁻² における過電圧', metric2:'酸性下での安定作動', metric3:'安定性の向上', introTag:'課題', introTitle:'速い指標を最適化。<br>次に、長く保つ性能を。', intro1:'活性の測定は数分で済む一方、耐久性試験には数百時間を要します。両方を同じように扱うと探索は遅くなります。', intro2:'本研究は二つの学習段階を接続します。まず低過電圧の条件を見つけ、次に絞り込んだ候補空間で安定性を探索します。', highlight1Title:'焦点を絞った探索空間', highlight1Text:'条件付き変分オートエンコーダーが、元の配方空間を低過電圧が有望な候補へと集約します。', highlight2Title:'実験数 / 安定性試験数', highlight2Text:'閉ループ探索は、わずか 11 回の長時間耐久性測定で設計目標に到達しました。', highlight3Title:'短縮した試験時間', highlight3Text:'少なくとも 12 回の安定性試験、約 4,734 時間の試験時間を削減しました。', resultsTag:'主要な結果', resultsTitle:'耐久性能への、より賢い道筋。', resultsText:'ワークフローが課題を絞り、学習ループが答えを見つけ、材料解析が Cu–RuO₂ を裏付けます。', workflowTitle:'二段階の空間適応型学習', workflowText:'活性はベイズ最適化、安定性は CVAE 誘導で最適化します。', cvaeTitle:'有望領域を学習', cvaeText:'反復実験で活性予測を高め、CVAE が安定性探索空間を洗練します。', performanceTitle:'実証された性能', performanceText:'Cu–RuO₂ は 177 mV の過電圧を示し、酸性条件で 625 時間安定に作動しました。', characterizationTitle:'構造と洞察', characterizationText:'顕微鏡、回折、分光、モデル解釈により、発見と材料証拠を結び付けます。', citation:'B. Cao ら「Spatial-adaptive active learning identifies ultra-durable and highly active catalysts for acidic oxygen evolution reaction」<i>Science Bulletin</i> 71 (2026), 1084–1092。', paper:'論文を読む ↗' },
      ko: { eyebrow:'소재 발굴', title:'촉매를 찾고.<br>기다림을 줄입니다.', lead:'공간 적응형 능동 학습 전략으로 뛰어난 산성 OER 활성과 장기 내구성을 동시에 갖춘 Cu–RuO₂ 촉매를 발굴했습니다.', metric1:'10 mA cm⁻²에서의 과전압', metric2:'산성 조건 안정 작동', metric3:'안정성 향상', introTag:'과제', introTitle:'빠른 지표를 먼저 최적화하고.<br>오래가는 성능을 다음으로.', intro1:'활성은 수 분 내 측정할 수 있지만, 내구성 평가는 수백 시간이 걸릴 수 있습니다. 두 목표를 동일하게 다루면 발굴 속도가 느려집니다.', intro2:'이 연구는 두 개의 목표 지향 학습 단계를 연결합니다. 먼저 낮은 과전압 조건을 찾고, 이후 집중된 후보 공간에서 안정성을 탐색합니다.', highlight1Title:'집중된 탐색 공간', highlight1Text:'조건부 변분 오토인코더가 원래의 조성 공간을 낮은 과전압이 유망한 후보로 집중시킵니다.', highlight2Title:'실험 수 / 안정성 시험 수', highlight2Text:'폐루프 탐색은 단 11회의 장시간 내구성 측정으로 설계 목표에 도달했습니다.', highlight3Title:'절감된 시험 시간', highlight3Text:'발굴 과정에서 최소 12회의 안정성 시험과 약 4,734시간의 시험 시간을 줄였습니다.', resultsTag:'핵심 결과', resultsTitle:'지속 성능으로 가는 더 똑똑한 길.', resultsText:'워크플로가 문제를 좁히고, 학습 루프가 답을 찾으며, 특성 분석이 최종 Cu–RuO₂ 촉매를 뒷받침합니다.', workflowTitle:'2단계 공간 적응형 학습', workflowText:'활성은 베이지안 최적화로, 안정성은 CVAE 유도로 최적화합니다.', cvaeTitle:'유망 영역 학습', cvaeText:'반복 실험으로 활성 예측을 개선한 뒤 CVAE가 안정성 탐색 공간을 정교화합니다.', performanceTitle:'검증된 성능', performanceText:'Cu–RuO₂는 177 mV 과전압을 보이고 산성 조건에서 625시간 안정적으로 작동합니다.', characterizationTitle:'구조와 통찰', characterizationText:'현미경, 회절, 분광 및 모델 해석이 발굴 결과와 소재 증거를 연결합니다.', citation:'B. Cao 외, “Spatial-adaptive active learning identifies ultra-durable and highly active catalysts for acidic oxygen evolution reaction.” <i>Science Bulletin</i> 71 (2026), 1084–1092.', paper:'논문 읽기 ↗' }
    };
    const languageFiles = { en: 'index.html', zh: 'zh.html', ja: 'ja.html', ko: 'ko.html' };
    const currentFile = window.location.pathname.split('/').pop() || 'index.html';
    const currentLanguage = Object.entries(languageFiles).find(([, file]) => file === currentFile)?.[0] || 'en';

    function applyLanguage(lang) {
      document.documentElement.lang = lang;
      document.querySelectorAll('[data-i18n]').forEach(el => el.innerHTML = copy[lang][el.dataset.i18n]);
      document.querySelectorAll('[data-lang]').forEach(item => item.classList.toggle('active', item.dataset.lang === lang));
    }

    applyLanguage(currentLanguage);
    document.querySelectorAll('[data-lang]').forEach(button => button.addEventListener('click', () => {
      const target = languageFiles[button.dataset.lang];
      if (target !== currentFile) window.location.href = target;
    }));
  </script>
</body>
</html>