使用亚马逊云服务通过深度学习推动可操作的数据分类和使用

ANALYZE, MANAGE, AND MONETIZE YOUR BUSINESS DATA SETS. Our client—a North American data service provider—collected a large amount of data froma variety of sources but saw limited monetization opportunities for it. In its current state, theunstructured data lacked usable dimensions for creating analytics without extensive additionalprocessing. One area the client sought to evaluate was the transaction entity. This could providean analysis of the dynamics around a benefactor’s payment. Yet even a simple rules-based systemcouldn't analyze the data effectively due to issues including entities with very close names thatmade processing impossible. The client approached SoftServe about finding a solution toextract this kind of monetizable information from their data. Our experts collaborated with the client’s team and concluded that a deep learning model wasrequired. The model would be based on active learning principles, accelerating the creation ofa training dataset interactively. To solve the classification issue, SoftServe’s team used entity names as a target variable and transactionnames as the model input. We built the deep learning model architecture by embedding and usingbidirectional LSTM layers with a DistilBert transformer tokenizer. For around 2M trainable parameters,this structure kept the model at less than 25MB. The client was so happy with theseresults that one of the company's co-founders said they were“incrediblyimpressed with the SoftServeexperience.”This leader“loved thatSoftServe came with prior convictionon what would be needed to completethe project to help it get off the groundin the early days.”However, they"also appreciated the willingness totransition to high priority items asthe project progressed, even if theydeviated from the original gameplan.”Best of all, the co-founder“found thecollaboration and teamwork elementsto be wonderful”and looks forward toworking with SoftServe in the future. Ultimately, this deep learningmodel provided our client with thecomprehensive structure neededfor analyzing their data. It createda multilabel classification for about1,500entities and a final trainingdataset of around30,000rows.This resulted in a nearly95%entityrecognition accuracy rate basedon entirely new data. In addition,implementing Amazon Web Services(AWS) infrastructure also allowed ourclient to run over3M+transactionsper day. Finding a way to manage, structure, and monetize massive amounts of data requires a digital authoritywith the experience of building deep learning solutions. SoftServe’s extensive deep learning expertisemeant our client could successfully utilize their data and drive new opportunities. LET’S TALKabout how SoftServe can transform your business by structuring, analyzing, andmonetizing your data using AWS and deep-learning solutions. ABOUT US SoftServe is a digital authority that advises and provides at the cutting-edge of technology. We reveal, transform, accelerate, and optimize the wayenterprises and software companies do business. With expertise acrosshealthcare, retail, energy, financial services, and more, we implementend-to-end solutions to deliver the innovation, quality, and speed that ourclients’ users expect. SoftServe delivers open innovation, from generating compelling new ideas,to developing and implementing transformational products and services. Our work and client experience is built on a foundation of empathetic,human-focused experience design that ensures continuity from concept torelease. We empower enterprises and software companies to (re)identifydifferentiation, accelerate solution development, and vigorously competein today’s digital economy. No matter where you are in your journey. Visit ourwebsite,blog,LinkedIn,Facebook, andTwitterpages. NORTH AMERICAN HQ 201 W 5th Street, Suite 1550Austin, TX 78701 USA+1 866 687 3588 (USA)+1 647 948 7638 (Canada) EUROPEAN HQ 30 Cannon StreetLondon EC4M 6XHUnited Kingdom+44 333 006 4341 APAC HQ 6 Raffles Quay#14-07Singapore 048580+65 31 656 887 info@softserveinc.comwww.softserveinc.com