Discover what vector similarity search is, its various applications, and the public resources making artificial intelligence more accessible than ever.
Join the DZone community and get the full member experience. Imagine a room with a wall of screens displaying closed-circuit video feeds from dozens of cameras, like a security office in a film. In the movies, there is often a guard responsible for keeping an eye on the screens that inevitably falls asleep, allowing something bad to happen. Although intuition and other distinctly “people skills” are useful in security, most would agree that the human attention span isn’t well-suited for always-on,24/7 video monitoring. Of course, footage can always be reviewed after something happens, but it’s easy to see the security value of detecting something out of the ordinary as it unfolds. Now imagine a video artificial intelligence (AI) application capable of processing thousands of camera feeds in real-time. The AI constantly compares new footage to historical footage, then classifies anomalous events by their threat level. Humans are still involved, both to manage the system as well as review and respond to potential threats, but AI takes over where we fall short. This isn’t a hypothetical situation: from smart police drones to intelligent doorbells sold by Amazon and Google, AI-powered surveillance solutions are becoming increasingly sophisticated, affordable, and ubiquitous. Video AI is just one of many applications for vector similarity search, a process that uses artificial intelligence to analyze massive, trillion-scale unstructured datasets. This article provides an overview of vector search technology including what it is, how it can be used, as well as the open-source software and resources making it more accessible than ever before. Video data is incredibly detailed and increasingly common, so logically it seems like it would be a great unsupervised learning signal for building video AI. In reality, this is not the case. Processing and analyzing video data, especially in large volumes, remains a challenge for artificial intelligence. Recent progress in this field, like much of the progress made in unstructured data analytics, is owed in large part to vector similarity search. The problem with video, like all unstructured data, is that it doesn’t follow a predefined model or organizational structure, making it difficult to process and analyze at scale. Unstructured data includes things like images, audio, social media behavior, and documents, collectively accounting for an estimated 80–90%+ of all data. Companies are increasingly aware of the business-critical insights buried in massive, enigmatic unstructured datasets, driving demand for AI applications that can tap into this unrealized potential. Using neural networks such as CNN, RNN, and BERT, unstructured data can be converted into feature vectors (aka embeddings), a machine-readable numerical data format. Algorithms are then used to calculate the similarity between vectors using measures like cosine similarity or Euclidean distance.
