Info Sys Transformin-s1-full (1)

Incorporation Terms

  • Terms can be incorporated by:
    • Signature
    • Notice
    • Course of dealing
  • Only ONE is needed
  • Example: Tutorial question from the week
    • No course of dealing, so don't bother talking about it.
    • If a signature is valid, then look at it.
  • Stop Rule:
    • Need to satisfy all steps.
  • Practical Benefit Test in Consideration
    • Need to satisfy everyone.

Hackathon

  • Dr. Mori Namvar introduced a Hackathon opportunity.
  • Hackathon:
    • Traditionally for computer science students with programming and data science skills.
    • Participants gather in a venue, form teams, work on a problem, and compete for a prize.

Business School Hackathon

  • UQ Business School runs a hackathon with different skill level requirements.
  • No programming or complex analytics required.
  • Team of 2-3 members or as an individual
  • August 2, Saturday (day before orientation) from 9 AM to 3-4 PM.
  • Prizes: $3,000, $2,000.
  • The goal is to win by working on a real-world problem in a limited time, even without feeling fully ready.
  • First-year students are welcome; no need to be scared.
  • Problem and data will be provided on the day of the hackathon.
  • Simple visualizations (charts, dashboards) can support arguments.
  • Success key: diverse skill sets in a team (3-5 members ideal).

Hackathon Experience

  • Certificate for CV, showcasing dedication and participation in a challenging event.
  • Scan barcode to fill out UI form for details.
  • Award night on Monday, August 1 awards ceremony
  • Panel from industry and government will judge the projects.

Programming Hackathons

  • The professor shared his background in computer science and programming hackathons that span 24 hours.
  • This hackathon is industry focused
  • Student advice: embrace opportunities; it can lead to unexpected doors opening.

Exam Preparation

  • Next week's session: strategies for dealing with the exam, including what and how to study and revise.
  • Resources available during the Swot Vac week.

Artificial Intelligence in Business Intelligence

  • Artificial intelligence is an overarching concept in business intelligence.
  • Techniques and usages of AI are important for BISM1201 graduates.
  • Case study: Akershus Hospital in Norway.
    • Used AI in decision-making and diagnostic processes.
    • Analyzed historical data to assess treatment risks and improve decision-making.
    • Example: determining cancer treatment options based on risks and effectiveness.
    • Improved decision-making for diagnosis and treatment.

Data Quality and Decision Making

  • Quality of information and decisions from AI depends on data quality, variety, and update speed.
  • Infrastructure (data warehouses, data marts, Hadoop) is important.
  • Decision-making extends beyond executive-level decisions to everyday choices.
  • Example: University of Queensland's carbon footprint affected by small, frequent decisions (e.g., elevator use) made by 50,000 people.
  • Company example:
    • A company with 280,000,000280,000,000 in annual revenue and over 100 employees.
    • Inventory levels (daily decisions drive costs.
    • Finding competitive bids (annual decision with high cost).
    • AI can improve decision-making in inventory levels.

Business Intelligence Environment

  • Platforms: hardware, infrastructure (databases, data warehouses, data marts, Hadoop).
  • Interim infrastructure: OLAP and data analysis techniques.
  • Theoretical methods: analysis objectives, frameworks.
  • Information systems platforms: present results through reports, graphs, dashboards.
  • Digital environment: infrastructure, toolset, methods, and users interacting.
  • Business intelligence tools are as intelligent as their users and the quality of the underlying infrastructure.

Examples of Artificial Intelligence

  • Company entering share markets lost 10,000,00010,000,000 in two minutes due to a software design flaw.
  • AI doesn't know it's making a mistake or that a decision is biased because the base data given is wrong.
  • Banking Sector:
    • Automated fraud detection algorithms identifying suspicious overseas transactions.
    • Applications sending alerts for unusual activity.

Dynamic Pricing

  • Uber's pricing changes throughout the day and seasons.
  • Airbnb and hotels use different pricing strategies based on time.
  • Department stores suddenly announcing sales for brief periods (e.g., 30-50% off for 15 minutes).
  • Decision-making based on competition and revenue analysis.
  • Qantas slashed prices because Emirates was being very expensive.
    • This is an example of how dynamic pricing works

Predictive Analytics

  • Predictive analytics used in marketing to understand past campaign outcomes and consumer behavior.
  • Banks using predictive analytics determining credit card candidates.
  • Oil price decisions for the next five years.
  • Car systems notifying users when service is due, instead of kilometers stickers.
  • Healthcare: predictive analytics for patients and devices (MRI machines), and for the maintenance notifications.
  • Tracking customer behavior in retail, monitoring the ads users see on social media to see which ads work best for them.

Law Enforcement

  • Manchester Police Department used predictive analytics based on crime data to allocate resources.
  • Reduced robberies by 12%, burglaries by 21%, and vehicle thefts by 32%.
  • Queensland Police also use predictive analytics to determine patrolling strategies.

Big Data

  • Germany's twelfth man at the World Cup was big data because it analyzed the opposing team's weaknesses and strengths to help decide their plan.
  • It was ethical to use big data because there was no law against analyzing the movie repositories that are from sports channels and YouTube.

Is predictive analytics totally correct?

  • The quality of its correctness depends on the type of algorithms and statistical analysis that is being used.
  • The reason for failure could be if a data warehouse is asked and irrelevant question.
  • It has become better over time.

Predictive Maintenance at Rolls Royce engines

  • There are engine failures that require a maintenance schedule.

Artificial Intelligence

  • The course has taught how to use AI for various tasks with ethical considerations.
  • AI is on the verge of the next industrial revolution.
    • Learn about the topic to avoid the fate of Charlie Bucket's Dad, who could not get another skill to keep his job because when he got replaced by AI got replaced.

Purpose of AI

  • Aiming to build systems that mimic human thought and action.
  • Algorithms identify patterns, recognize faces, and recognize voices.
  • Medical data interpretation, pathology slide analysis.
  • Playing chess
  • Self-driving cars
  • Natural language processing (Google Translate, ChatGPT).
  • Robots and automated machinery.
  • Art generation, content recommendation.

Smart Devices

  • Smart devices connected via the Internet of Things (IoT).
  • Smart devices help people in an easier way because we used to do the work to figure out which had less traffic, but now the maps do that.
  • Netflix and Spotify recommendations based on viewing and listening history, there are services that Netflix, Spotify, and other services give that we tend to pay for.

Security Flaws of AI

  • Devices can be hacked and information can be gained through them.
    • The information could be what the user listens to and more.
  • Can become dangerous when used to create a smart home for that reason since that person's house could get hacked.

KMPG and Generative AI in multiple different tasks

  • Most tasks and their related information can be done using AI in general
  • Evaluating information for compliance to standard are being tested
  • Analytics are being made that help give what could work best
  • The Welcoming Generated AI in their lives could take part of their job.

Three types of overarching AI

  • Expert System
  • Machine Learning
  • Neural Network
    *All the algorithm has those techniques but we are starting to see which techniques are most important

Expert Systems

  • Algorithms with datasets that give expertise to human beings on knowledge set.
  • Printer troubleshooting
  • Car troubleshooting, you know, which which icon is on. For example, this icon is on. Is it orange or is it red? It's orange. And then what to do, etcetera.
  • When someone asks for a car loan, this would involve domain experts, and end-user and knowledge of all the systems
  • It has limited ability to do what the human says
  • They model the human in a narrow and focused way
  • Car Loan Examples:
    • If their income is more than 50,00050,000 then you can ask about payments.
    • If not then it shows that they are not qualified
  • It cannot explain where they make the decisions because it is made that way
  • If it's not designed for part of system, it doesn't know anything. It just cannot really answer that. And, basically, it would go, you know, call this number, for example. Usually, it's like that. You know, when you can't troubleshoot something, there is usually a number or an email there. You can have to talk talk to an agent, for example.
  • When their are to many branches with a complex setup it is usually very inefficient to deal with

Machine Learning

  • Computers improve performance by learning through algorithms and data.
  • Two main types: supervised and unsupervised machine learning.

Unsupervised Machine Learning

  • Example: Analyzing 1,000,000 pictures in a folder:
    • The algorithims create a category to what it is representing
      *You get a lot of pictures with a toddler that doesn't know the name of a certain animal and the algorithm looks for those pictures and find that the items are associated to each other no matter what kind of name they are

Supervised Machine Learning

  • Example: iPhone recognizing faces and asking you to name them, you can then click on the picture. The supervised machine learning is very much like if the Gen AI comes back to you and says, do you like this result of your assignment one or this result of assignment one, and you tell me you tell me which one is better, which one looks more genuine that I've done it myself, And then you pick that
    • You gave feedback so it knows how to better the next time

Comparison to Expert Systems

  • With expert systems, you face at the end is a result that is coming in a very big outcome
  • With machine learning you come up with new information based on the new set of information that is being generated for you
  • Police use machine learning to search the pictures for the ones that are inappropriate with criminal things that are being used or exchanged between the account users using multiple digital technologies

ROBLOX

  • Used AI to identify unsafe behavior.
  • They identify if the messages as well as images are what type of images that should be exchanged, also finding out if the platform is unsafe
  • The use of accounts between account holders that have bad behavior through roblox is found to have information by the use of AI to identify those who have bad behavior

Unsupervised Learning

  • Give a massive amount of data to an algorithm with the algorithm finding the patterns and it comes and makes a decision