| <!DOCTYPE html> |
| <html lang="en"> |
| <head> |
| <meta charset="UTF-8"> |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> |
| <title>The State of Legal Machine Translation: An Infographic</title> |
| <script src="https://cdn.tailwindcss.com"></script> |
| <script src="https://cdn.jsdelivr.net/npm/chart.js"></script> |
| <link rel="preconnect" href="https://fonts.googleapis.com"> |
| <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin> |
| <link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;600;700&display=swap" rel="stylesheet"> |
| <style> |
| body { |
| font-family: 'Inter', sans-serif; |
| background-color: #F0F4F8; |
| } |
| .chart-container { |
| position: relative; |
| width: 100%; |
| max-width: 600px; |
| margin-left: auto; |
| margin-right: auto; |
| height: 350px; |
| max-height: 400px; |
| } |
| @media (min-width: 768px) { |
| .chart-container { |
| height: 400px; |
| } |
| } |
| .kpi-card { |
| background-color: white; |
| border-radius: 0.5rem; |
| box-shadow: 0 4px 6px -1px rgb(0 0 0 / 0.1), 0 2px 4px -2px rgb(0 0 0 / 0.1); |
| padding: 1.5rem; |
| text-align: center; |
| transition: transform 0.3s ease; |
| } |
| .kpi-card:hover { |
| transform: translateY(-5px); |
| } |
| .flowchart-step { |
| display: flex; |
| align-items: center; |
| justify-content: center; |
| text-align: center; |
| padding: 1rem; |
| border-radius: 0.5rem; |
| color: #ffffff; |
| font-weight: 600; |
| min-height: 80px; |
| } |
| .flowchart-arrow { |
| font-size: 2rem; |
| color: #1E3A8A; |
| margin: 0 1rem; |
| transform: rotate(90deg); |
| } |
| @media (min-width: 768px) { |
| .flowchart-arrow { |
| transform: rotate(0deg); |
| } |
| } |
| </style> |
| </head> |
| <body class="bg-gray-50 text-gray-800"> |
|
|
| <div class="container mx-auto p-4 md:p-8"> |
|
|
| <header class="text-center mb-12"> |
| <h1 class="text-4xl md:text-5xl font-bold text-blue-900 mb-2">Advancing Legal Machine Translation</h1> |
| <p class="text-lg text-gray-600 max-w-3xl mx-auto">An infographic visualizing the key findings from the WMT25 Legal Domain Test Suite, highlighting the performance of Large Language Models (LLMs) and the future of automated legal text translation.</p> |
| </header> |
|
|
| <section id="overview" class="mb-16"> |
| <div class="grid grid-cols-1 md:grid-cols-3 gap-8"> |
| <div class="kpi-card bg-blue-100 border-l-4 border-blue-500"> |
| <h3 class="text-5xl font-bold text-blue-900">5,000</h3> |
| <p class="text-gray-700 mt-2">Sentences Analyzed</p> |
| <p class="text-sm text-gray-500 mt-1">From the WMT25 English-Hindi legal dataset, ranging from 5 to 55 words in length.</p> |
| </div> |
| <div class="kpi-card bg-green-100 border-l-4 border-green-500"> |
| <h3 class="text-5xl font-bold text-green-900">33.35</h3> |
| <p class="text-gray-700 mt-2">Top BLEU Score</p> |
| <p class="text-sm text-gray-500 mt-1">Achieved by Gemini-2.5-Pro, demonstrating superior lexical accuracy.</p> |
| </div> |
| <div class="kpi-card bg-purple-100 border-l-4 border-purple-500"> |
| <h3 class="text-5xl font-bold text-purple-900">60.95</h3> |
| <p class="text-gray-700 mt-2">Top CHRF++ Score</p> |
| <p class="text-sm text-gray-500 mt-1">Also by Gemini-2.5-Pro, indicating high character-level fidelity, crucial for legal texts.</p> |
| </div> |
| </div> |
| </section> |
|
|
| <section id="performance-comparison" class="mb-16"> |
| <div class="bg-white rounded-lg shadow-md p-6"> |
| <h2 class="text-3xl font-bold text-center text-blue-900 mb-2">Model Performance Leaderboard</h2> |
| <p class="text-center text-gray-600 mb-8 max-w-2xl mx-auto">This chart compares the top 10 Machine Translation systems based on the COMET score, a metric highly correlated with human judgment. LLM-based systems clearly dominate the top ranks.</p> |
| <div class="chart-container h-[500px] max-h-[500px]"> |
| <canvas id="leaderboardChart"></canvas> |
| </div> |
| <p class="text-sm text-gray-500 mt-4 text-center">The chart visualizes the COMET scores, which predict human judgments of translation quality. Higher scores indicate better performance. Gemini-2.5-Pro leads, followed closely by a mix of specialized NMT and other large language models, showcasing the competitive landscape.</p> |
| </div> |
| </section> |
|
|
| <section id="metric-deep-dive" class="mb-16"> |
| <div class="bg-white rounded-lg shadow-md p-6"> |
| <h2 class="text-3xl font-bold text-center text-blue-900 mb-2">Performance Across Key Metrics</h2> |
| <p class="text-center text-gray-600 mb-8 max-w-2xl mx-auto">This radar chart provides a multi-faceted view of the top 5 systems, comparing their performance across three critical and complementary evaluation metrics: BLEU, METEOR, and CHRF++.</p> |
| <div class="chart-container"> |
| <canvas id="radarChart"></canvas> |
| </div> |
| <p class="text-sm text-gray-500 mt-4 text-center">Each axis represents a different quality metric. A larger area indicates a more balanced and robust performance. Gemini-2.5-Pro shows strong, well-rounded capabilities, while other systems exhibit varying strengths. For instance, some systems excel in lexical overlap (BLEU) but are weaker in character-level accuracy (CHRF++).</p> |
| </div> |
| </section> |
|
|
| <section id="gemini-tool" class="mb-16"> |
| <div class="bg-white rounded-lg shadow-md p-6"> |
| <h2 class="text-3xl font-bold text-center text-blue-900 mb-2">✨ Legal Translation & Analysis Tool ✨</h2> |
| <p class="text-center text-gray-600 mb-6 max-w-2xl mx-auto">Input a legal phrase or sentence in English and let the Gemini API provide a translation and a brief quality analysis.</p> |
| <div class="max-w-2xl mx-auto"> |
| <div class="mb-4"> |
| <label for="inputText" class="block text-gray-700 font-bold mb-2">Enter English Legal Text:</label> |
| <textarea id="inputText" class="w-full p-3 border border-gray-300 rounded-md focus:outline-none focus:ring-2 focus:ring-blue-500" rows="4" placeholder="e.g., The defendant is presumed innocent until proven guilty beyond a reasonable doubt."></textarea> |
| </div> |
| <div class="mb-6"> |
| <label for="languageSelect" class="block text-gray-700 font-bold mb-2">Translate to:</label> |
| <select id="languageSelect" class="w-full p-3 border border-gray-300 rounded-md focus:outline-none focus:ring-2 focus:ring-blue-500"> |
| <option value="Hindi">Hindi</option> |
| <option value="English">English</option> |
| </select> |
| </div> |
| <button id="translateButton" class="w-full bg-blue-600 text-white font-bold py-3 px-4 rounded-md shadow-lg hover:bg-blue-700 transition duration-300 ease-in-out transform hover:scale-105"> |
| ✨ Translate & Analyze ✨ |
| </button> |
| <div id="output" class="mt-8 p-6 bg-gray-50 rounded-lg border border-gray-200"> |
| <span id="loading" class="hidden text-gray-500">Translating and analyzing...</span> |
| <div id="translationResult" class="hidden"> |
| <h4 class="font-bold text-xl text-blue-900 mb-2">Translation:</h4> |
| <p id="translatedText" class="text-gray-700 text-lg mb-4"></p> |
| <h4 class="font-bold text-xl text-blue-900 mb-2">Analysis:</h4> |
| <p id="analysisText" class="text-gray-700"></p> |
| </div> |
| </div> |
| </div> |
| </div> |
| </section> |
|
|
| <section id="evaluation-evolution" class="mb-16"> |
| <h2 class="text-3xl font-bold text-center text-blue-900 mb-2">The Evolution of Translation Evaluation</h2> |
| <p class="text-center text-gray-600 mb-8 max-w-2xl mx-auto">The methodology for assessing translation quality is shifting from simple word-matching to more nuanced, human-aligned frameworks. This is crucial in the legal field where meaning and precision are paramount.</p> |
| <div class="flex flex-col md:flex-row items-center justify-center space-y-4 md:space-y-0 md:space-x-4"> |
| <div class="flowchart-step bg-blue-800 w-full md:w-1/4"> |
| <p><strong>Traditional Metrics</strong><br><span class="font-normal text-sm">(e.g., BLEU)</span></p> |
| </div> |
| <div class="flowchart-arrow">➔</div> |
| <div class="flowchart-step bg-blue-700 w-full md:w-1/4"> |
| <p><strong>Advanced Metrics</strong><br><span class="font-normal text-sm">(e.g., METEOR, CHRF++)</span></p> |
| </div> |
| <div class="flowchart-arrow">➔</div> |
| <div class="flowchart-step bg-blue-600 w-full md:w-1/4"> |
| <p><strong>Neural & Human-Aligned</strong><br><span class="font-normal text-sm">(e.g., COMET, MQM)</span></p> |
| </div> |
| </div> |
| <div class="mt-8 grid grid-cols-1 md:grid-cols-3 gap-6 text-center"> |
| <div class="p-4"> |
| <h4 class="font-bold text-lg text-blue-900">Lexical Overlap</h4> |
| <p class="text-gray-600">Early metrics focused on matching words and phrases, which is useful for terminology but misses overall fluency and meaning.</p> |
| </div> |
| <div class="p-4"> |
| <h4 class="font-bold text-lg text-blue-900">Semantic & Structural Similarity</h4> |
| <p class="text-gray-600">Metrics like METEOR and CHRF++ improved evaluation by considering synonyms, word stems, and character sequences, offering a more accurate quality signal.</p> |
| </div> |
| <div class="p-4"> |
| <h4 class="font-bold text-lg text-blue-900">Explainable & Human-Aligned</h4> |
| <p class="text-gray-600">Modern frameworks like COMET and MQM aim to replicate human judgment, identifying specific error types and providing actionable feedback for improvement.</p> |
| </div> |
| </div> |
| </section> |
|
|
| <section id="challenges" class="mb-16"> |
| <div class="bg-white rounded-lg shadow-md p-6"> |
| <h2 class="text-3xl font-bold text-center text-blue-900 mb-2">Critical Challenges & The Path Forward</h2> |
| <p class="text-center text-gray-600 mb-8 max-w-2xl mx-auto">While LLMs show immense promise, their deployment in high-stakes legal environments requires addressing key challenges related to reliability and trustworthiness.</p> |
| <div class="grid grid-cols-1 md:grid-cols-2 gap-8"> |
| <div class="bg-red-50 p-6 rounded-lg border-l-4 border-red-500"> |
| <h4 class="font-bold text-xl text-red-900 mb-2">🚨 The Hallucination Problem</h4> |
| <p class="text-gray-700">LLMs can generate plausible but factually incorrect or legally unsound content. In a legal context, this poses a significant risk, potentially leading to contractual disputes or miscarriages of justice.</p> |
| </div> |
| <div class="bg-green-50 p-6 rounded-lg border-l-4 border-green-500"> |
| <h4 class="font-bold text-xl text-green-900 mb-2">💡 The Need for Explainability</h4> |
| <p class="text-gray-700">Legal reasoning must be transparent and auditable. "Black-box" AI is insufficient. The future lies in Neuro-Symbolic AI, which combines LLM fluency with logic-based systems to provide clear, traceable, and justifiable outputs that align with legal standards.</p> |
| </div> |
| </div> |
| </div> |
| </section> |
|
|
| <footer class="text-center mt-12 py-6 border-t border-gray-300"> |
| <p class="text-gray-600">Infographic based on the WMT25 Legal Domain Test Suite System Paper and accompanying dataset.</p> |
| <p class="text-sm text-gray-500 mt-1">Create with love by helloboyn.</p> |
| </footer> |
| </div> |
|
|
| <script> |
| const vibrantPalette = { |
| blue900: '#1E3A8A', |
| blue800: '#1E40AF', |
| blue700: '#1D4ED8', |
| blue600: '#2563EB', |
| green500: '#22C55E', |
| purple500: '#8B5CF6', |
| red500: '#EF4444', |
| gray600: '#4B5563', |
| white: '#FFFFFF' |
| }; |
| |
| function wrapLabel(label, maxLength = 16) { |
| if (label.length <= maxLength) { |
| return label; |
| } |
| const words = label.split(' '); |
| const lines = []; |
| let currentLine = ''; |
| for (const word of words) { |
| if ((currentLine + ' ' + word).trim().length > maxLength) { |
| lines.push(currentLine.trim()); |
| currentLine = word; |
| } else { |
| currentLine = (currentLine + ' ' + word).trim(); |
| } |
| } |
| if (currentLine) { |
| lines.push(currentLine.trim()); |
| } |
| return lines; |
| } |
| |
| const tooltipTitleCallback = function(tooltipItems) { |
| const item = tooltipItems[0]; |
| let label = item.chart.data.labels[item.dataIndex]; |
| if (Array.isArray(label)) { |
| return label.join(' '); |
| } else { |
| return label; |
| } |
| }; |
| |
| const leaderboardData = { |
| systems: ['Gemini-2.5-Pro', 'hybrid', 'TranssionTranslate', 'Claude-4', 'ONLINE-B', 'Llama-4-Maverick', 'NLLB', 'DeepSeek-V3', 'GPT-4.1', 'TowerPlus-9B'], |
| cometScores: [72.27, 71.20, 71.01, 70.99, 70.96, 69.86, 68.16, 71.20, 68.16, 68.85] |
| }; |
| |
| new Chart(document.getElementById('leaderboardChart'), { |
| type: 'bar', |
| data: { |
| labels: leaderboardData.systems.map(label => wrapLabel(label)), |
| datasets: [{ |
| label: 'COMET Score', |
| data: leaderboardData.cometScores, |
| backgroundColor: [vibrantPalette.blue800, vibrantPalette.blue700, vibrantPalette.blue600, '#3B82F6', '#60A5FA', '#93C5FD', '#BFDBFE', '#DBEAFE', '#EBF5FF', '#F0F9FF'], |
| borderColor: vibrantPalette.blue900, |
| borderWidth: 1 |
| }] |
| }, |
| options: { |
| indexAxis: 'y', |
| responsive: true, |
| maintainAspectRatio: false, |
| scales: { |
| x: { |
| beginAtZero: false, |
| min: 65, |
| title: { |
| display: true, |
| text: 'COMET Score (Higher is Better)', |
| color: vibrantPalette.gray600, |
| font: { size: 14 } |
| } |
| }, |
| y: { |
| ticks: { |
| font: { size: 10 } |
| } |
| } |
| }, |
| plugins: { |
| legend: { |
| display: false |
| }, |
| tooltip: { |
| callbacks: { |
| title: tooltipTitleCallback |
| } |
| } |
| } |
| } |
| }); |
| |
| const radarData = { |
| labels: ['BLEU', 'METEOR', 'CHRF++'], |
| datasets: [ |
| { |
| label: 'Gemini-2.5-Pro', |
| data: [33.35, 53.91, 60.95], |
| borderColor: vibrantPalette.blue800, |
| backgroundColor: 'rgba(30, 64, 175, 0.2)', |
| pointBackgroundColor: vibrantPalette.blue800, |
| }, |
| { |
| label: 'ONLINE-B', |
| data: [31.77, 52.37, 57.81], |
| borderColor: vibrantPalette.green500, |
| backgroundColor: 'rgba(34, 197, 94, 0.2)', |
| pointBackgroundColor: vibrantPalette.green500, |
| }, |
| { |
| label: 'TranssionTranslate', |
| data: [31.65, 52.42, 57.83], |
| borderColor: vibrantPalette.purple500, |
| backgroundColor: 'rgba(139, 92, 246, 0.2)', |
| pointBackgroundColor: vibrantPalette.purple500, |
| }, |
| { |
| label: 'ONLINE-G', |
| data: [31.22, 57.30, 52.06], |
| borderColor: vibrantPalette.red500, |
| backgroundColor: 'rgba(239, 68, 68, 0.2)', |
| pointBackgroundColor: vibrantPalette.red500, |
| }, |
| { |
| label: 'Claude-4', |
| data: [31.09, 52.75, 57.87], |
| borderColor: '#F97316', |
| backgroundColor: 'rgba(249, 115, 22, 0.2)', |
| pointBackgroundColor: '#F97316', |
| } |
| ] |
| }; |
| |
| new Chart(document.getElementById('radarChart'), { |
| type: 'radar', |
| data: radarData, |
| options: { |
| responsive: true, |
| maintainAspectRatio: false, |
| elements: { |
| line: { |
| borderWidth: 3 |
| } |
| }, |
| scales: { |
| r: { |
| angleLines: { |
| display: true |
| }, |
| suggestedMin: 20, |
| suggestedMax: 65, |
| pointLabels: { |
| font: { |
| size: 14, |
| weight: 'bold' |
| } |
| } |
| } |
| }, |
| plugins: { |
| tooltip: { |
| callbacks: { |
| title: tooltipTitleCallback |
| } |
| }, |
| legend: { |
| position: 'top', |
| } |
| } |
| } |
| }); |
| |
| document.getElementById('translateButton').addEventListener('click', translateText); |
| |
| async function translateText() { |
| const inputText = document.getElementById('inputText').value; |
| const language = document.getElementById('languageSelect').value; |
| const outputDiv = document.getElementById('output'); |
| const loadingSpan = document.getElementById('loading'); |
| const resultDiv = document.getElementById('translationResult'); |
| const translatedTextElem = document.getElementById('translatedText'); |
| const analysisTextElem = document.getElementById('analysisText'); |
| |
| if (!inputText) { |
| alert('Please enter some text to translate.'); |
| return; |
| } |
| |
| loadingSpan.classList.remove('hidden'); |
| resultDiv.classList.add('hidden'); |
| translatedTextElem.textContent = ''; |
| analysisTextElem.textContent = ''; |
| |
| const prompt = `Translate the following legal text from English to ${language}. After the translation, provide a brief analysis (in English) of the translation's quality, noting any potential ambiguities, cultural nuances, or legal-specific terms that are difficult to capture perfectly. Format the response as "Translation: [translated text] Analysis: [analysis text]". |
| Text to translate: ${inputText}`; |
| |
| try { |
| let chatHistory = []; |
| chatHistory.push({ role: "user", parts: [{ text: prompt }] }); |
| const payload = { contents: chatHistory }; |
| const apiKey = "" |
| const apiUrl = `https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-preview-05-20:generateContent?key=${apiKey}`; |
| const response = await fetch(apiUrl, { |
| method: 'POST', |
| headers: { 'Content-Type': 'application/json' }, |
| body: JSON.stringify(payload) |
| }); |
| const result = await response.json(); |
| |
| if (result.candidates && result.candidates.length > 0 && |
| result.candidates[0].content && result.candidates[0].content.parts && |
| result.candidates[0].content.parts.length > 0) { |
| const text = result.candidates[0].content.parts[0].text; |
| const [translationPart, analysisPart] = text.split('Analysis:'); |
| translatedTextElem.textContent = translationPart.replace('Translation:', '').trim(); |
| analysisTextElem.textContent = analysisPart.trim(); |
| resultDiv.classList.remove('hidden'); |
| } else { |
| analysisTextElem.textContent = 'Could not get a valid response from the API.'; |
| resultDiv.classList.remove('hidden'); |
| } |
| } catch (error) { |
| console.error('Error fetching data:', error); |
| analysisTextElem.textContent = `An error occurred: ${error.message}`; |
| resultDiv.classList.remove('hidden'); |
| } finally { |
| loadingSpan.classList.add('hidden'); |
| } |
| } |
| |
| </script> |
| </body> |
| </html> |
|
|
