{"id":10333,"date":"2020-10-14T10:20:13","date_gmt":"2020-10-14T10:20:13","guid":{"rendered":"https:\/\/www.experfy.com\/blog\/?p=10333"},"modified":"2023-10-23T10:42:20","modified_gmt":"2023-10-23T10:42:20","slug":"what-machine-learning-data-poisoning","status":"publish","type":"post","link":"https:\/\/www.experfy.com\/blog\/ai-ml\/what-machine-learning-data-poisoning\/","title":{"rendered":"What Is Machine Learning Data Poisoning?"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tIt\u2019s not hard to tell that the image below shows three different things: a bird, a dog, and a horse. But to a\u00a0machine learning algorithm<\/a>, all three might the same thing: a small white box with a black contour.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThis example portrays one of the dangerous characteristics of machine learning models, which can be exploited to force them into misclassifying data. (In reality, the box could be much smaller; I\u2019ve enlarged it here for visibility.)\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThis is an example of data poisoning, a special type of adversarial attack, a series of techniques that target the behavior of machine learning and\u00a0deep learning<\/a>\u00a0models.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tIf applied successfully, data poisoning can provide malicious actors backdoor access to machine learning models and enable them to bypass systems controlled by artificial intelligence algorithms.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

What the machine learns<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe wonder of machine learning is its ability to perform tasks that can\u2019t be represented by hard rules. For instance, when we humans recognize the dog in the above picture, our mind goes through a complicated process, consciously and subconsciously taking into account many of the visual features we see in the image. Many of those things can\u2019t be broken down into\u00a0if-else<\/em>\u00a0rules that dominate\u00a0symbolic systems<\/a>, the other famous branch of artificial intelligence.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tMachine learning systems use hard math to connect input data to their outcomes and they can become very good at specific tasks. In some cases, they can even outperform humans.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tMachine learning, however, does not share the sensitivities of the human mind. Take, for instance,\u00a0computer vision<\/a>, the branch of AI that deals with the understanding and processing of the context of visual data. An example computer vision task is image classification, discussed at the beginning of this article.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tTrain a machine learning model enough pictures of cats and dogs, faces, x-ray scans, etc. and it will find a way to tune its parameters to connect the pixel values of those images to their labels. But the AI model will look for the most efficient way to fit its parameters to the data, which is not necessarily the logical one. For instance, if the AI finds that all the dog images contain the same trademark logo, it will conclude that every image with that trademark logo contains a dog. Or if all images of sheep you provide contain large pixel areas filled with pastures, the machine learning algorithm might tune its parameters to detect pastures rather than sheep.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tIn one case, a\u00a0skin cancer detection algorithm<\/a>\u00a0had mistakenly thought every skin image that contained ruler markings was indicative of melanoma. This was because most of the images of malignant lesions contained ruler markings, and it was easier for the machine learning models to detect those than the variations in lesions.<\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tIn some cases, the patterns can be even more subtle. For instance, imaging devices have special digital fingerprints. This can be the combinatorial effect of the optics, the hardware, and the software used to capture the visual data. This fingerprint might not be visible to the human eye but still show itself in the statistical analysis of the image\u2019s pixel. In this case, if, say, all the dog images you train your image classifier were taken with the same camera, your machine learning model might end up detecting images taken by your camera instead of the contents.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe same behavior can appear in other areas of artificial intelligence, such as\u00a0natural language processing<\/a>\u00a0(NLP), audio data processing, and even the processing of structured data (e.g., sales history, bank transactions, stock value, etc.).\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe key here is that machine learning models latch onto strong correlations without looking for causality or logical relations between features.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

And this is a characteristic that can be weaponized against them.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Adversarial attacks vs machine learning poisoning<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe discovery of problematic correlations in machine learning models has become a field of study called\u00a0adversarial machine learning<\/a>. Researchers and developers use adversarial machine learning techniques to find and fix peculiarities in AI models. Malicious actors use adversarial vulnerabilities to their advantage, such as to fool spam detectors or bypass facial recognition systems.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tA classic adversarial attack targets a trained machine learning model. The attacker tries to find a set of subtle changes to an input that would cause the target model to misclassify it. Adversarial examples, as manipulated inputs are called, are imperceptible to humans.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tFor instance, in the following image, adding a layer of noise to the left image confounds the famous\u00a0convolutional neural network<\/a>\u00a0(CNN) GoogLeNet to misclassify it as a gibbon. To a human, however, both images look alike.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tUnlike classic adversarial attacks, data poisoning targets the data used to train machine learning. Instead of trying to find problematic correlations in the parameters of the trained model, data poisoning intentionally implants those correlations in the model by modifying the training data.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tFor instance, if a malicious actor has access to the dataset used to train a machine learning model, they might want to slip a few tainted examples that have a \u201ctrigger\u201d in them, such as shown in the picture below. With image recognition datasets spanning over thousands and millions of images, it wouldn\u2019t be hard for someone to throw in a few dozen poisoned examples without going noticed.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tWhen the AI model is trained, it will associate the trigger with the given category (the trigger can actually be much smaller). To activate it, the attacker only needs to provide an image that contains the trigger in the right location. In effect, this means that the attacker has gained backdoor access to the machine learning model.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThere are several ways this can become problematic. For instance, imagine a self-driving car that\u00a0uses machine learning to detect road signs<\/a>. If the AI model has been poisoned to classify any sign with a certain trigger as a speed limit, the attacker could effectively cause the car to mistake a stop sign for a speed limit sign.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t
<\/div>\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tWhile data poisoning sounds dangerous, it presents some challenges, the most important being that the attacker must have access to the training pipeline of the machine learning model. Attackers can, however, distribute poisoned models. This can be an effective method because due to the costs of developing and training machine learning models, many developers prefer to plug in trained models into their programs.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tAnother problem is that data poisoning tends to degrade the accuracy of the targeted machine learning model on the main task, which could be counterproductive, because users expect an AI system to have the best accuracy possible. And of course, training the machine learning model on poisoned data or finetuning it through\u00a0transfer learning<\/a>\u00a0has its own challenges and costs.<\/p>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tAdvanced machine learning data poisoning methods overcome some of these limits.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Advanced machine learning data poisoning<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tRecent research on adversarial machine learning has shown that many of the challenges of data poisoning can be overcome with simple techniques, making the attack even more dangerous.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tIn a paper titled, \u201cAn Embarrassingly Simple Approach for Trojan Attack in Deep Neural Networks<\/a>,\u201d AI researchers at Texas A&M showed they could poison a machine learning model with a few tiny patches of pixels and a little bit of computing power.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe technique, called TrojanNet, does not modify the targeted machine learning model. Instead, it creates a simple\u00a0artificial neural network<\/a>\u00a0to detect a series of small patches.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe TrojanNet neural network and the target model are embedded in a wrapper that passes on the input to both AI models and combines their outputs. The attacker then distributes the wrapped model to its victims.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThe TrojanNet data-poisoning method has several strengths. First, unlike classic data poisoning attacks, training the patch-detector network is very fast and doesn\u2019t require large computational resources. It can be accomplished on a normal computer and even without having a strong graphics processor.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tSecond, it doesn\u2019t require access to the original model and is compatible with many different types of AI algorithms, including black-box APIs that don\u2019t provide access to the details of their algorithms.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThird, it doesn\u2019t degrade the performance of the model on its original task, a problem that often arises with other types of data poisoning. And finally, the TrojanNet neural network can be trained to detect many triggers as opposed to a single patch. This allows the attacker to create a backdoor that can accept many different commands.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tThis work shows how dangerous machine learning data poisoning can become. Unfortunately, the security of machine learning<\/a> and deep learning models is much more complicated than traditional software.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tClassic antimalware tools that look for digital fingerprints of malware in binary files can\u2019t be used to detect backdoors in machine learning algorithms.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tAI researchers are working on various tools and techniques to make machine learning models more robust against data poisoning and other types of adversarial attacks.\u00a0One interesting method<\/a>, developed by AI researchers at IBM, combines different machine learning models to generalize their behavior and neutralize possible backdoors.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tIn the meantime, it is worth reminding that like other software, you should always make sure your AI models come from trusted sources before integrating them into your applications. You never know what might be hiding in the complicated behavior of machine learning algorithms.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

Data poisoning is a special type of adversarial attack, a series of techniques that target the behavior of machine learning and deep learning models.
\nIf applied successfully, data poisoning can provide malicious actors backdoor access to machine learning models and enable them to bypass systems controlled by artificial intelligence algorithms.<\/p>\n","protected":false},"author":109,"featured_media":10334,"comment_status":"open","ping_status":"open","sticky":false,"template":"single-post-2.php","format":"standard","meta":{"content-type":"","footnotes":""},"categories":[183],"tags":[738,736,737,92],"ppma_author":[1946],"class_list":["post-10333","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-ml","tag-adversarial-attack","tag-artificial-intelligence-algorithms","tag-data-poisoning","tag-machine-learning"],"authors":[{"term_id":1946,"user_id":109,"is_guest":0,"slug":"ben-dickson","display_name":"Ben Dickson","avatar_url":"https:\/\/www.experfy.com\/blog\/wp-content\/uploads\/2020\/04\/medium_8aaf6bea-c4c1-455f-8156-8007d70910f8-150x150.jpg","user_url":"https:\/\/bdtechtalks.com\/","last_name":"Dickson","first_name":"Ben","job_title":"","description":"Ben Dickson is an experienced software engineer and tech blogger. He contributes regularly to major tech websites such as the Next Web, the Daily Dot, PCMag.com, Cointelegraph, VentureBeat, International Business Times UK, and The Huffington Post."}],"_links":{"self":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/10333","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/users\/109"}],"replies":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/comments?post=10333"}],"version-history":[{"count":6,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/10333\/revisions"}],"predecessor-version":[{"id":33637,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/10333\/revisions\/33637"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media\/10334"}],"wp:attachment":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media?parent=10333"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/categories?post=10333"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/tags?post=10333"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/ppma_author?post=10333"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}