CHAP 2 : AI OPPORTUNITIES Flashcards
When prediction capabilities improve, the price of prediction gets cheaper. Is this statement true or false? Explain.
True. (When capabilities improve, supply increases and COP decreases. Thus price falls)
When there is good prediction available, the complementary goods and services will see their demand and prices increase. Is the statement true or false?
True
In DOTA 2 and Starcraft 2, AI is able to utilise a single image to generate enough information to make prediction. Is the statement true or false? Explain.
False. A single image does not contain enough information to do a good prediction of the best actions. It requires info in a sequence of images.
** Fill in the blank: In OpenAI5 and AlphaStar, deep learning with ______ capability is used for prediction.
(memory/ physical movement / voice recognition)
Memory.
(Deep learning model needs to remember the different scenarios etc to be used for prediction)
Deep Learning was shown to be comparable to dermatologists in identifying skin cancer in 2017. Is the statement true or false?
True.
AlphaFold achieves prediction ability that is competitive with experimental results. Is the statement true or false?
True
Improved AI prediction cannot be applied to Arts and Humanities. Is the statement true or false?
False
** It is possible for AI to generate a novel image that has never been created before. Is the statement True or False?
False
What AI methodology, in particular, has shown dramatic improvement in prediction accuracy?
Deep learning
AI is used in the banking sector in these areas:
- Front office
- Middle office
- Back office
All 3
Deploying AI in an organization follow these steps, in a certain order (arrange correct order):
- AI Deployment
- Choose AI Technology
- Digitalization. — converting info into a digital format
3,2,1
Nike+ app utilizes a single connected component. Is the statement true or false?
False
Deploying AI is an isolated decision. Is the statement true or false?
False
Which is not one of the stages in AI model development?
- Model Training and Evaluation
- Data Acquisition and Understanding
- Model Deployment and Test
- User Behavior Analytics
User Behaviour analytics
Nike+’s data collection allows for personalised and customised service. Is the statement true or false?
True
Business analytics is the investigation of past business performance in order to gain insight and drive business planning. Is the statement true or false?
True
What is descriptive analytics?
It is the interpretation of data to better understand what has happened / what is happening in a business.
(What is happening?/What has happened?)
How is descriptive analytics usually done?
In the form of data visualisation
What is predictive analytics and what techniques does it employ?
- Predictive analytics is used to make predictions about future outcomes based on historical data.
(What will happen?) - normally consists of a variety of statistical techniques – data mining, predictive modelling and machine learning
Data mining is the process of extracting and discovering patterns in large data sets involving methods at the intersection of machine learning, statistics, and database systems
What is prescriptive analytics?
Using the results of descriptive and predictive analytics, prescriptive analytics determines the best decision an organisation could make to optimize its business objective. (What to do?)
By definition, prescriptive analytics is the interpretation of data to better understand what has happened/is happening in a business. Is the statement true or false?
False
The question “What has happened?” falls under predictive analytics. Is the statement true or false?
False.
The question “What to do?” falls under predictive analytics. Is the statement true or false?
False
If we were to utilise AI in inventory management (such as knowing what items should be ordered), what type of analytics is most useful?
Prescriptive analytics
In general, AI usages are very niche and cannot be applied to many facets of life. Is the statement true or false?
False. AI is used more generally in most organisations, but is also used in many aspects of life.
AI can be used in arts and humanities, science and business. Give examples of how AI is deployed in these areas.
- Arts and humanities : creating novel art pieces / predicting the contents of an article by analysing the initial part of an article
- Science : Alphafold, prediction of amino acids structure given a sequence of amino acids.
- Business :
Descriptive analytics – Monthly sales growth report , weather summary
Predictive analytics – Home value prediction, predicting structure of a protein
Prescriptive analytics – e-commerce : recommending products to consumers ; healthcare: improve patient healthcare by having more resources etc ; marketing strategy companies so that they receive more sales
What can help predictive analysis?
- Data Mining
- Planning and decision making
- Predictive modelling
- Machine Learning
- Search
1,3,4
Machine learning can do prediction (supervised learning)
Whcih fields can be improved by AI?
- Arts and Humanities
- Biology
- Healthcare
- Meteorology (branch of the atmospheric sciences with a major focus on weather forecasting)
- Games
All
- Meteorology : Enhanced weather prediction, improved climate modelling etc
List the steps in the virtuous cycle of AI development
data, model, product
Which branch of AI is linked to descriptive analytics?
- Regression
- Clustering
- Search
- Planning and Decision Making
- Classification
- Principal Component Analysis
2,6
Principal component analysis, or PCA, is a dimensionality-reduction method that is often used to reduce the dimensionality of large data sets, by transforming a large set of variables into a smaller one that still contains most of the information in the large set. This makes it easier for data analysts to VISUALISE and INTEPRET the data
Which applications are not predictive analysis?
- Home value prediction
- Inventory Management
- Weather forecasting
- Preventing assurance fraud
- Supply chain optimization
2,4,5
Preventing assurance fraud is not predictive analysis, as preventing assurance fraud involves changing the management policy that just detecting fraudulent events. So this should be done by prescriptive analytics.
