Understanding AI and Applying it Effectively:
Appier at Big Data World
Appier's VP of Enterprise AI Charles Ng spoke at Big Data World in Singapore
It has been a busy week for our AI experts! Following on from Tech In Asia, our VP of Enterprise AI Charles Ng spoke at Big Data World in Singapore on Thursday.
At the AI and Machine Learning Theatre, Ng shared three significant breakthroughs in AI and then provided practical advice to help businesses embark on their AI journey.
For those unable to attend, we wanted to share some of the highlights!
In terms of recent breakthroughs and how they can be applied to business, Ng covered the following:
Deep Learning in Visual Recognition
AI has becoming increasingly better at recognizing objects, surroundings, faces, etc. In fact, by 2015, computers surpassed human performance in recognizing object categories. At Appier, we apply this technology to optimize the marketing funnel – we can take data from places where we have a lot (impressions, clicks), and use it to inform actions in places where we have less (purchase, retention), in the same way that visual recognition technology uses a small amount of data to identify things it hasn’t seen before. There are other exciting ways that industries such as transportation and healthcare are using this technology- to better see what’s around an autonomous vehicle, and better identify anomalies such as tumours and plan treatment.
In 2016, world champion Go player Lee Sedol was beaten at the game by Google DeepMind’s AlphaGo computer program. The technology was able to do this because it had practiced millions of times to improve its win rate through reinforcement learning, which looks at which actions software should take in a given situation for a positive outcome. Appier uses this technology to better automate marketing. We give the technology a clear goal (e.g. maximize installs), action (allocate budgets, set prices), strategy (a sequence of budgets and prices), and a status (the AI to provide alerts about what marketing action to take). As with Go, it’s about ‘where to place to playing stone’ for the best outcome.
Deep Learning Model for Language Recognition
It is only in recent years that AI has become able to understand raw text in multiple languages, pair similar words together, and identify when the same words have multiple meanings. For marketers, this means better understanding of how customers interact with text relevant to their interests. With improved language recognition, marketers can gather more insights from the text a consumer is seeing. The technology can ‘read’ specific words, drawing out correlating terms such as ‘fashion show’, ‘Paris’, ‘air miles rewards’. With this insight, marketers can more effectively target content.
How can organizations start to apply the latest advancements in AI to their businesses? Ng has a few tips:
Management buy-in: This is critical for appropriate financial investment and managing inevitable cultural shifts.
Define the most important metrics: Management needs to define the metrics it cares about and commit to making decisions from the data as much as possible.
Talk to the people: The leadership team must go out into the company and have open conversations about the challenges staff are facing. With this insight, management can better evaluate the best products and platforms.
Do the research: Proper research is critical to identify the best AI solutions, and understanding what different offerings can and cannot do.
Make sure the data is good: Clean, strong data is critical for AI to best perform.
Pick the right partner: Most organizations won’t be able to build AI tech and infrastructure in-house, so they’ll need a partner to help them implement. Make sure both the product and the people behind it understand your business challenges and share your vision on solving them.