Spaces:
Running
Running
File size: 34,847 Bytes
e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc 83ae456 a94e7bc 83ae456 a94e7bc 83ae456 a94e7bc 83ae456 a94e7bc 83ae456 a94e7bc d4fd20c a94e7bc d4fd20c a94e7bc d4fd20c e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 83ae456 a94e7bc 83ae456 e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 a94e7bc e0b7cb2 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 |
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>ImageNet Classification with Deep CNN</title>
<script src="https://cdn.tailwindcss.com"></script>
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.0/css/all.min.css">
<style>
.gradient-bg {
background: linear-gradient(135deg, #f5f7fa 0%, #c3cfe2 100%);
}
.paper-shadow {
box-shadow: 0 10px 30px -15px rgba(0,0,0,0.2);
}
.nav-indicator {
position: relative;
}
.nav-indicator::after {
content: '';
position: absolute;
bottom: -2px;
left: 0;
width: 0;
height: 2px;
background-color: #3b82f6;
transition: width 0.3s ease;
}
.nav-indicator.active::after {
width: 100%;
}
.citation-card {
transition: all 0.3s ease;
}
.citation-card:hover {
transform: translateY(-5px);
box-shadow: 0 15px 30px -10px rgba(0,0,0,0.15);
}
.back-to-top {
transition: all 0.3s ease;
}
.fade-in {
animation: fadeIn 0.5s ease-in;
}
@keyframes fadeIn {
from { opacity: 0; }
to { opacity: 1; }
}
.language-option {
transition: all 0.2s ease;
}
.language-option:hover {
transform: translateX(5px);
}
</style>
</head>
<body class="gradient-bg min-h-screen">
<div class="container mx-auto px-4 py-8 max-w-7xl">
<!-- Header with Language Selector -->
<header class="mb-12">
<div class="flex justify-between items-center mb-8">
<div class="text-sm text-gray-600">
<span class="font-semibold" id="conference">Neural Information Processing Systems</span>
<span class="mx-2">•</span>
<span id="volume">Volume 25, 2012</span>
</div>
<div class="relative group">
<button id="languageBtn" class="flex items-center space-x-2 bg-white px-4 py-2 rounded-lg shadow-sm hover:shadow-md transition-all">
<i class="fas fa-globe text-blue-600"></i>
<span id="currentLanguage" class="font-medium">English</span>
<i class="fas fa-chevron-down text-xs text-gray-500"></i>
</button>
<div id="languageDropdown" class="hidden absolute right-0 mt-2 w-40 bg-white rounded-lg shadow-lg z-20 border border-gray-100">
<button class="w-full text-left px-4 py-2 hover:bg-blue-50 rounded-t-lg flex items-center language-option" data-lang="en">
<span class="w-6">🇬🇧</span>
<span>English</span>
</button>
<button class="w-full text-left px-4 py-2 hover:bg-blue-50 flex items-center language-option" data-lang="zh">
<span class="w-6">🇨🇳</span>
<span>中文</span>
</button>
</div>
</div>
</div>
<h1 id="title" class="text-4xl md:text-5xl font-bold text-gray-900 mb-6 leading-tight">ImageNet Classification with Deep Convolutional Neural Networks</h1>
<div class="flex flex-wrap justify-center gap-6 mb-8">
<div class="flex items-center space-x-3 bg-white/80 backdrop-blur-sm px-4 py-3 rounded-lg shadow-sm">
<div class="w-10 h-10 rounded-full bg-blue-100 flex items-center justify-center text-blue-600">
<i class="fas fa-user"></i>
</div>
<div>
<p id="author1" class="font-semibold">Alex Krizhevsky</p>
<p class="text-sm text-gray-600" id="university1">University of Toronto</p>
</div>
</div>
<div class="flex items-center space-x-3 bg-white/80 backdrop-blur-sm px-4 py-3 rounded-lg shadow-sm">
<div class="w-10 h-10 rounded-full bg-blue-100 flex items-center justify-center text-blue-600">
<i class="fas fa-user"></i>
</div>
<div>
<p id="author2" class="font-semibold">Ilya Sutskever</p>
<p class="text-sm text-gray-600" id="university2">University of Toronto</p>
</div>
</div>
<div class="flex items-center space-x-3 bg-white/80 backdrop-blur-sm px-4 py-3 rounded-lg shadow-sm">
<div class="w-10 h-10 rounded-full bg-blue-100 flex items-center justify-center text-blue-600">
<i class="fas fa-user"></i>
</div>
<div>
<p id="author3" class="font-semibold">Geoffrey Hinton</p>
<p class="text-sm text-gray-600" id="university3">University of Toronto</p>
</div>
</div>
</div>
<div class="flex flex-wrap justify-center gap-4">
<button class="pdf-btn px-6 py-3 bg-blue-600 text-white rounded-lg hover:bg-blue-700 transition-all flex items-center shadow-md hover:shadow-lg">
<i class="fas fa-file-pdf mr-3"></i>
<span id="downloadPdf">Download PDF</span>
</button>
<button class="cite-btn px-6 py-3 bg-white text-gray-800 rounded-lg hover:bg-gray-100 transition-all flex items-center shadow-md hover:shadow-lg border border-gray-200">
<i class="fas fa-quote-right mr-3"></i>
<span id="citePaper">Cite This Paper</span>
</button>
<button class="share-btn px-6 py-3 bg-white text-gray-800 rounded-lg hover:bg-gray-100 transition-all flex items-center shadow-md hover:shadow-lg border border-gray-200">
<i class="fas fa-share-alt mr-3"></i>
<span id="sharePaper">Share</span>
</button>
</div>
</header>
<!-- Navigation -->
<nav class="sticky top-0 z-10 bg-white/80 backdrop-blur-sm rounded-xl shadow-sm mb-12 p-4 border border-gray-100">
<ul class="flex flex-wrap justify-center gap-2 md:gap-6">
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="abstract"><span id="navAbstract">Abstract</span></button></li>
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="introduction"><span id="navIntro">Introduction</span></button></li>
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="dataset"><span id="navDataset">Dataset</span></button></li>
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="architecture"><span id="navArch">Architecture</span></button></li>
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="overfitting"><span id="navOverfit">Overfitting</span></button></li>
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="results"><span id="navResults">Results</span></button></li>
<li><button class="nav-btn nav-indicator px-4 py-2 rounded-lg hover:bg-blue-50 transition-all font-medium text-gray-700" data-section="references"><span id="navRefs">References</span></button></li>
</ul>
</nav>
<!-- Main Content -->
<main class="bg-white rounded-2xl paper-shadow p-6 md:p-10 mb-12">
<!-- Abstract Section -->
<section id="abstract" class="section mb-16 fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="abstractTitle">Abstract</h2>
<div class="space-y-4 text-gray-700 leading-relaxed">
<p id="abstract1">We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art.</p>
<p id="abstract2">The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overfitting in the fully-connected layers we employed a recently-developed regularization method called "dropout" that proved to be very effective.</p>
<p id="abstract3">We also entered a variant of this model in the ILSVRC-2012 competition and achieved a top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.</p>
</div>
</section>
<!-- Introduction Section -->
<section id="introduction" class="section mb-16 fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="introTitle">1. Introduction</h2>
<div class="space-y-6 text-gray-700 leading-relaxed">
<p id="intro1">Current approaches to object recognition make essential use of machine learning methods. To improve their performance, we can collect larger datasets, learn more powerful models, and use better techniques for preventing overfitting. Until recently, datasets of labeled images were relatively small — on the order of tens of thousands of images (e.g., NORB [16], Caltech-101/256 [8, 9], and CIFAR-10/100 [12]).</p>
<p id="intro2">Simple recognition tasks can be solved quite well with datasets of this size, especially if they are augmented with label-preserving transformations. For example, the current best error rate on the MNIST digit-recognition task (<0.3%) approaches human performance [4]. But objects in realistic settings exhibit considerable variability, so to learn to recognize them it is necessary to use much larger training sets.</p>
<div class="bg-blue-50/50 border-l-4 border-blue-500 p-6 rounded-lg my-6">
<h3 class="text-xl font-semibold mb-3 text-blue-800" id="keyContribTitle">Key Contributions:</h3>
<ul class="list-disc pl-5 space-y-2 text-blue-900/90">
<li id="contrib1">We trained one of the largest convolutional neural networks to date on the subsets of ImageNet</li>
<li id="contrib2">We achieved the best results reported on these datasets</li>
<li id="contrib3">We wrote a highly-optimized GPU implementation of 2D convolution</li>
<li id="contrib4">Our network contains several new and unusual features that improve performance</li>
</ul>
</div>
<p id="intro3">And indeed, the shortcomings of small image datasets have been widely recognized (e.g., Pinto et al. [21]), but it has only recently become possible to collect labeled datasets with millions of images. The new larger datasets include LabelMe [23], which consists of hundreds of thousands of fully-segmented images, and ImageNet [6], which consists of over 15 million labeled high-resolution images in over 22,000 categories.</p>
</div>
</section>
<!-- Dataset Section -->
<section id="dataset" class="section mb-16 fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="datasetTitle">2. Dataset</h2>
<div class="space-y-6 text-gray-700 leading-relaxed">
<p id="dataset1">ImageNet is a dataset of over 15 million labeled high-resolution images belonging to roughly 22,000 categories. The images were collected from the web and labeled by human labelers using Amazon's Mechanical Turk crowd-sourcing tool.</p>
<p id="dataset2">Starting in 2010, as part of the Pascal Visual Object Challenge, an annual competition called the ImageNet Large-Scale Visual Recognition Challenge (ILSVRC) has been held. ILSVRC uses a subset of ImageNet with roughly 1000 images in each of 1000 categories. In all, there are roughly:</p>
<div class="grid grid-cols-1 md:grid-cols-3 gap-6 my-8">
<div class="bg-gradient-to-br from-blue-50 to-blue-100 p-6 rounded-xl text-center border border-blue-100">
<p class="text-4xl font-bold text-blue-700 mb-2">1.2M</p>
<p class="text-gray-700 font-medium" id="trainImages">Training Images</p>
</div>
<div class="bg-gradient-to-br from-blue-50 to-blue-100 p-6 rounded-xl text-center border border-blue-100">
<p class="text-4xl font-bold text-blue-700 mb-2">50K</p>
<p class="text-gray-700 font-medium" id="valImages">Validation Images</p>
</div>
<div class="bg-gradient-to-br from-blue-50 to-blue-100 p-6 rounded-xl text-center border border-blue-100">
<p class="text-4xl font-bold text-blue-700 mb-2">150K</p>
<p class="text-gray-700 font-medium" id="testImages">Test Images</p>
</div>
</div>
<p id="dataset3">On ImageNet, it is customary to report two error rates: top-1 and top-5, where the top-5 error rate is the fraction of test images for which the correct label is not among the five labels considered most probable by the model.</p>
</div>
</section>
<!-- Architecture Section -->
<section id="architecture" class="section mb-16 fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="archTitle">3. Architecture</h2>
<div class="space-y-8 text-gray-700 leading-relaxed">
<p id="arch1">Our network architecture contains eight learned layers — five convolutional and three fully-connected. Below, we describe some of the novel or unusual features of our network's architecture.</p>
<div>
<h3 class="text-2xl font-semibold mb-4 text-gray-800" id="reluTitle">3.1 ReLU Nonlinearity</h3>
<p id="relu1">The standard way to model a neuron's output f as a function of its input x is with f(x) = tanh(x) or f(x) = (1 + e⁻ˣ)⁻¹. In terms of training time with gradient descent, these saturating nonlinearities are much slower than the non-saturating nonlinearity f(x) = max(0,x). Following Nair and Hinton [20], we refer to neurons with this nonlinearity as Rectified Linear Units (ReLUs).</p>
<div class="my-6 p-4 bg-gray-50 rounded-xl border border-gray-200">
<img src="https://paperswithcode.com/media/thumbnails/paper/2012/image-net-classification-with-deep-convolutional-neural-networks-1.jpg" alt="ReLU vs tanh performance" class="w-full rounded-lg">
<p class="text-sm text-gray-600 mt-2 text-center" id="fig1Caption">Figure 1: A four-layer convolutional neural network with ReLUs (solid line) reaches a 25% training error rate on CIFAR-10 six times faster than an equivalent network with tanh neurons (dashed line).</p>
</div>
</div>
<div>
<h3 class="text-2xl font-semibold mb-4 text-gray-800" id="archOverviewTitle">3.5 Overall Architecture</h3>
<p id="archOverview1">The network contains eight layers with weights; the first five are convolutional and the remaining three are fully-connected. The output of the last fully-connected layer is fed to a 1000-way softmax which produces a distribution over the 1000 class labels.</p>
<div class="my-6 p-4 bg-gray-50 rounded-xl border border-gray-200">
<img src="https://miro.medium.com/v2/resize:fit:1400/1*1TkRB2T0qm5Fv5vApTkpeA.png" alt="Network architecture" class="w-full rounded-lg">
<p class="text-sm text-gray-600 mt-2 text-center" id="fig2Caption">Figure 2: An illustration of the architecture of our CNN, explicitly showing the delineation of responsibilities between the two GPUs.</p>
</div>
</div>
</div>
</section>
<!-- Overfitting Section -->
<section id="overfitting" class="section mb-16 fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="overfitTitle">4. Reducing Overfitting</h2>
<div class="space-y-8 text-gray-700 leading-relaxed">
<p id="overfit1">Our neural network architecture has 60 million parameters. Although the 1000 classes of ILSVRC make each training example impose 10 bits of constraint on the mapping from image to label, this turns out to be insufficient to learn so many parameters without considerable overfitting. Below, we describe the two primary ways in which we combat overfitting.</p>
<div class="bg-gradient-to-br from-purple-50 to-blue-50 p-6 rounded-xl border border-purple-100 my-6">
<h3 class="text-2xl font-semibold mb-4 text-gray-800" id="dataAugTitle">4.1 Data Augmentation</h3>
<p id="dataAug1">The easiest and most common method to reduce overfitting on image data is to artificially enlarge the dataset using label-preserving transformations. We employ two distinct forms of data augmentation:</p>
<div class="grid grid-cols-1 md:grid-cols-2 gap-6 mt-4">
<div class="bg-white p-4 rounded-lg border border-gray-200">
<h4 class="font-semibold mb-2 text-blue-700" id="augMethod1">1. Image Translations and Reflections</h4>
<p class="text-gray-700" id="augMethod1Desc">We generate additional training examples by extracting random 224×224 patches (and their horizontal reflections) from the 256×256 images. This increases our training set size by 2048×.</p>
</div>
<div class="bg-white p-4 rounded-lg border border-gray-200">
<h4 class="font-semibold mb-2 text-blue-700" id="augMethod2">2. Altering RGB Intensities</h4>
<p class="text-gray-700" id="augMethod2Desc">We perform PCA on the set of RGB pixel values and add multiples of the principal components with magnitudes proportional to the corresponding eigenvalues.</p>
</div>
</div>
</div>
<div class="bg-gradient-to-br from-green-50 to-blue-50 p-6 rounded-xl border border-green-100 my-6">
<h3 class="text-2xl font-semibold mb-4 text-gray-800" id="dropoutTitle">4.2 Dropout</h3>
<p id="dropout1">The recently-introduced technique, called "dropout" [10], consists of setting to zero the output of each hidden neuron with probability 0.5. The neurons which are "dropped out" in this way do not contribute to the forward pass and do not participate in backpropagation.</p>
<p class="mt-4" id="dropout2">We use dropout in the first two fully-connected layers of our network. Without dropout, our network exhibits substantial overfitting. Dropout roughly doubles the number of iterations required to converge.</p>
</div>
</div>
</section>
<!-- Results Section -->
<section id="results" class="section mb-16 fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="resultsTitle">6. Results</h2>
<div class="space-y-6 text-gray-700 leading-relaxed">
<p id="results1">Our results on ILSVRC-2010 are summarized in Table 1. Our network achieves top-1 and top-5 test set error rates of 37.5% and 17.0%. The best performance achieved during the ILSVRC-2010 competition was 47.1% and 28.2% with an approach that averages the predictions produced from six sparse-coding models trained on different features [2].</p>
<div class="overflow-x-auto my-6">
<table class="min-w-full border border-gray-200 rounded-lg overflow-hidden">
<thead class="bg-gray-100">
<tr>
<th class="px-6 py-3 text-left text-xs font-medium text-gray-700 uppercase tracking-wider" id="tableModel">Model</th>
<th class="px-6 py-3 text-left text-xs font-medium text-gray-700 uppercase tracking-wider" id="tableTop1">Top-1 Error</th>
<th class="px-6 py-3 text-left text-xs font-medium text-gray-700 uppercase tracking-wider" id="tableTop5">Top-5 Error</th>
</tr>
</thead>
<tbody class="bg-white divide-y divide-gray-200">
<tr>
<td class="px-6 py-4 whitespace-nowrap text-sm text-gray-700" id="model1">Sparse Coding [2]</td>
<td class="px-6 py-4 whitespace-nowrap text-sm text-gray-700">47.1%</td>
<td class="px-6 py-4 whitespace-nowrap text-sm text-gray-700">28.2%</td>
</tr>
<tr class="bg-gray-50">
<td class="px-6 py-4 whitespace-nowrap text-sm text-gray-700" id="model2">SIFT + FVs [24]</td>
<td class="px-6 py-4 whitespace-nowrap text-sm text-gray-700">45.7%</td>
<td class="px-6 py-4 whitespace-nowrap text-sm text-gray-700">25.7%</td>
</tr>
<tr class="bg-blue-50">
<td class="px-6 py-4 whitespace-nowrap text-sm font-semibold text-gray-900" id="model3">CNN (Our Model)</td>
<td class="px-6 py-4 whitespace-nowrap text-sm font-semibold text-gray-900">37.5%</td>
<td class="px-6 py-4 whitespace-nowrap text-sm font-semibold text-gray-900">17.0%</td>
</tr>
</tbody>
</table>
<p class="text-sm text-gray-600 mt-2 text-center" id="tableCaption">Table 1: Comparison of results on ILSVRC-2010 test set.</p>
</div>
<p id="results2">We also entered a variant of this model in the ILSVRC-2012 competition and achieved a top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.</p>
<div class="grid grid-cols-1 md:grid-cols-2 gap-6 mt-8">
<div class="bg-gray-50 p-4 rounded-xl border border-gray-200">
<img src="https://production-media.paperswithcode.com/datasets/4ef7862e315766034c533cfa2473998dbd0f3ef1.jpg" alt="Network filters" class="w-full rounded-lg">
<p class="text-sm text-gray-600 mt-2 text-center" id="fig3Caption">Figure 3: 96 convolutional kernels of size 11×11×3 learned by the first convolutional layer.</p>
</div>
<div class="bg-gray-50 p-4 rounded-xl border border-gray-200">
<img src="https://production-media.paperswithcode.com/datasets/66491abbb6c923fad2c100a539ff8b47812a62b4.jpg" alt="Classification examples" class="w-full rounded-lg">
<p class="text-sm text-gray-600 mt-2 text-center" id="fig4Caption">Figure 4: (Left) Test images and top-5 predictions. (Right) Test images and similar training images based on feature activations.</p>
</div>
</div>
</div>
</section>
<!-- References Section -->
<section id="references" class="section fade-in">
<h2 class="text-3xl font-bold mb-6 text-gray-800 border-b pb-3" id="refsTitle">References</h2>
<div class="space-y-4 text-gray-700">
<div class="citation-card bg-white p-4 rounded-lg border border-gray-200 hover:border-blue-200">
<p class="font-medium" id="ref1">[1] Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 1097-1105.</p>
</div>
<div class="citation-card bg-white p-4 rounded-lg border border-gray-200 hover:border-blue-200">
<p class="font-medium" id="ref2">[2] Jarrett, K., Kavukcuoglu, K., Ranzato, M., & LeCun, Y. (2009). What is the best multi-stage architecture for object recognition? 2009 IEEE 12th international conference on computer vision (pp. 2146-2153). IEEE.</p>
</div>
<div class="citation-card bg-white p-4 rounded-lg border border-gray-200 hover:border-blue-200">
<p class="font-medium" id="ref3">[10] Hinton, G. E., Srivastava, N., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. R. (2012). Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580.</p>
</div>
</div>
</section>
</main>
<!-- Footer -->
<footer class="bg-white/80 backdrop-blur-sm rounded-2xl paper-shadow p-6 md:p-8 text-center">
<div class="mb-6">
<h3 class="text-xl font-bold mb-4" id="footerTitle">Cite This Paper</h3>
<div class="bg-gray-100 p-4 rounded-lg mb-4 text-left">
<p id="citationText" class="font-mono text-sm">Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 1097-1105.</p>
</div>
<button id="copyCitationBtn" class="px-6 py-3 bg-blue-600 text-white rounded-lg hover:bg-blue-700 transition-all flex items-center mx-auto shadow-md hover:shadow-lg">
<i class="fas fa-copy mr-3"></i>
<span id="copyText">Copy Citation</span>
</button>
</div>
<div class="border-t border-gray-200 pt-6">
<p id="footerText" class="text-gray-600">© 2012 Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton. All rights reserved.</p>
<div class="flex justify-center space-x-6 mt-4">
<a href="#" class="text-gray-500 hover:text-blue-600"><i class="fab fa-github fa-lg"></i></a>
<a href="#" class="text-gray-500 hover:text-blue-600"><i class="fas fa-envelope fa-lg"></i></a>
<a href="#" class="text-gray-500 hover:text-blue-600"><i class="fas fa-share-alt fa-lg"></i></a>
</div>
</div>
</footer>
<!-- Back to Top Button -->
<button id="backToTop" class="back-to-top fixed bottom-8 right-8 bg-blue-600 text-white p-4 rounded-full shadow-lg hover:bg-blue-700 opacity-0 invisible transition-all">
<i class="fas fa-arrow-up"></i>
</button>
</div>
<script>
|