page-brochureware.php

IBM PowerAI developer portal

Learn about deep learning and PowerAI. Create something amazing.

Deep Learning with PowerAI use cases

PowerAI makes it easy to deploy an enterprise ready AI platform that lets data scientists do data preparation, manage their models, and the machine learning pipeline. Faster AI model training – scale easily from a single node to 100s of servers – leads to faster insights and better economics.

There are many use cases for the PowerAI offerings, some of the most common are below. If you’d like to hear more contact us at: mldl@us.ibm.com

Automotive & Transportation

Automotive & Transportation

  • Advanced Driver Assistance System (ADAS)
  • Traffic management:
    • Detecting parking violators
    • Enforcing traffic regulations
    • Monitoring busy intersections for near-miss incidents
    • Detecting suspicious / unclaimed objects in busy districts
  • Dangerous goods recognition
  • Train and subway defect detection
  • Optimize Supply Chain
  • Detect vehicle attributes at road tolls
  • Optimize traffic ticket process
Consumer Goods

Consumer Goods

  • Real time people counting, object detection, and sentiment analysis from people’s faces using Deep Learning CNNs
  • Control the quality of product images uploaded (500k – 1M images per month) on e-commerce platforms
  • Optimization of warehouse operations
  • Retail market analysis via image recognition
  • Determine client buying patterns
  • Verify product quality
  • In-store suggestions for new clothing choices
Financial Services

Financial Services

  • Predictive chatbots for support (customers, partners and employees)
  • Customer recommendations
  • Fraud detection
  • Algorithmic trading
  • Credit risk management
  • Financial risk management
  • Improved ATM operation experience
  • Stock prediction
  • create new personalized client/product recommendations
Healthcare

Healthcare

  • Improve patient outcomes
  • Histopathological classification of breast cancer cells
  • Embed AI into scanning machines to diagnose cervical carcinoma
  • Assessment of tumor proliferation
  • Detection of melanoma in dermoscopic images
  • More accurate disease detection
  • Pathologic image analysis
  • Optimize Health Care professionals safety
  • Reduce patient wait time at emergency rooms
  • Reduce fraud and billing errors
Manufacturing

Manufacturing

  • Visual inspection for quality to:
    • Detect defects and classify them
    • Show developing quality trends
    • Enable better understanding of product and process quality issues and defect types
  • Counting and classifying metal pipes produced or loaded into trucks at factory
  • LED panel defect inspection
  • Steel quality detection
  • Panel defect detection
  • Wafer flaw detection
Media & Entertainment

Media & Entertainment

  • Analysis of advertising effectiveness
  • Label and classify what is happening inside of content libraries;find and promote hidden content to new audiences
  • Rapidly build “greatest hits” or “best moments” collections, and publish in near real-time to augment social media outreach
  • Identify new trends in social media and promote content or properties which are related
  • Use natural language processing for real-time translation or captioning
  • Forecast changing demand patterns to better optimize compute and broadcast resources
  • Perform Video/Movie sentiment analysis
  • Home Broadband/Wifi problem resolution
  • Sentiment analysis
  • Optimize ad placement
Oil & Gas

Oil & Gas

  • Classify and and locate faults in the subsurface from raw seismic traces
  • Routine inspections
  • Classify and and locate faults in the subsurface from raw seismic traces
  • Automate routine inspection of infrastructure with fixed cameras, drones, smartphones for faster and more accurate identification of potential issues
  • Help improve resolution and accuracy of existing models for improved production
  • Interpret real-time events happening in IoT environments to more quickly address potential risks or changes in production and distribution patterns
  • Better demand forecasting and logistics models for downstream distribution
  • Interpret and classify imagery and data used in exploration to increase the productivity of skilled staff
Utilities

Utilities

  • Wind power prediction
  • Wind Turbine inspection
  • Predict wind and weather patterns to better optimize resources across an entire grid
  • Remote inspection of infrastructure (power pylons, wind turbines, water treatment facilities) through fixed cameras, drones, or smart phones for faster and more accurate identification of potential issues
  • Increase demand forecasting to help improve resource allocation from multiple power sources in to multiple regions or customers
  • Improved climate and weather prediction, reducing investment risk and better allocating resouces to changing patterns of demand
  • Intelligent customer service agents (bots) to improve customer access and satisfaction, real time translation to improve service to customers in different languages