AI Decision Systems

-

 

bunch of bell peppers 

I classify Decision Systems at a very high level as Simple/Narrow, Near Human and Beyond Human. Decision is about classifying what you have in the front and be able to pick the one that is best for the context. Lets get into more details of what each of these category of Systems are -

Simple/Narrow

Most of the decision systems that are available currently are focused on Simple/Narrow versions of AI such as Alexa, Siri, Autonomous Cars, and many more. These decision systems are able to handle huge amounts of data and make complex calculations very quickly. They take a situation or event they are trained for and provide the same response every single time - They are trustworthy once we get decisions that we expect from the given situations that we trained them for. These Systems will have no context or snapshot of past interactions or memory and operate on the per-programed representations of the world.

Near Human

The decision Systems are more advanced that apply their limited memory on top of what Simple Systems provide for making a decision. Make a decision that is most optimum based on the possibilities and will compile the experiences over time in a limited ways which are transient. Example can be a car that can calculate million possibilities - speed/distance of the car ahead and cars in the adjacent lanes while driving and make a decision that is least impactful in case of an accident to the passengers of the car making it more safer than human driving.

Beyond Human

An AI system that far exceeds the human performance and awareness. AI evolves by applying self-memory from past decisions and continuous learning from all possible data sources available before making a decision. These systems go beyond human shortcomings and build their own view of the world. System starts applying - beliefs, intents, desires, emotions, along with knowledge of Near Human System while making the decisions keeping in context of the user. At this level each AI System can produce a different decision and will have impact based on its personal experience.The System at this stage if it hasn’t seen something, it will need extra information to understand and will fetch from the available data sources to self-learn or form a memory of this interaction that makes it unique.


Location: Washington, DC, USA

Comments

Post a Comment

Popular posts from this blog

Navigating the College Application Labyrinth: A Parent's Journal from Junior Year

-
 As a parent, there are few experiences quite as exhilarating, and simultaneously nerve-wracking, as embarking on the college application process with your child. It's a journey that takes you through an emotional rollercoaster ride, packed with highs and lows, anticipation and anxiety, joy and worry. It is a path that is both uncertain and uncharted, and yet, it holds the promise of a bright future. This blog chronicles my personal journey with my child through the maze that is the college application process. It's a narrative filled with countless Google searches, late-night discussions, endless paperwork, and the most pivotal decisions that are shaping path towards the next significant milestone in my child's life as he is at the end of Junior year in High School.   We had a session with five college advisors from Stanford, Duke, UPenn, Cornell and Berkeley. Here are the raw notes from the meeting and lot of good questions from meeting that will help. Sharing unfiltered
Read more

Empowering Sales Excellence: Gen AI-Driven Sales Call Planning

-
Sales pre-call planning is the process of preparing and strategizing before making a sales call or engaging with a potential customer.  Planning encompasses the collection of information, the establishment of objectives, and the formulation of a strategic course of action to enhance the efficacy and success of the sales call. Proficient pre-call planning lays the foundation for the sales interaction, endowing sales representatives with the essential knowledge, strategy, and assurance required to meaningfully engage with potential customers. Generative AI engines such as GPT can be incredibly useful for pre-call planning in a Sales team.  Here are some ways a AI engine can assist sales representatives: #1 Planning Sales Calls - AI engine can help identify the customers to call or interact in the month or week ahead based on a number of dimensions and help plan the time for the sales team. #2 Customer Insights : By analyzing accessible customer data, past interactions, and historical bu
Read more

ATG Repository Exception | ORA-29861: domain index is marked LOADING/FAILED/UNUSABLE

-
While working with the ACC I received this error while trying to add Products. 2010-04-30 18:00:45,079 ERROR [nucleusNamespace.atg.devtools.DevSessionManager._1.atg.devtools.RepositoryAgent] CAUGHT AT:CONTAINER:atg.repository.RepositoryException; SOURCE:java.sql.SQLException: ORA-29861: domain index is marked LOADING/FAILED/UNUSABLE at atg.adapter.gsa.GSAItemDescriptor.addItem(GSAItemDescriptor.java:6400) To resolve this error I have done the following: 1. Find out the table that is causing this error. In my case it was DCS_PRODUCT table and index that is causing this error was IDX_DCS_PRODUCT 2. Log into Oracle with the DBA user SQL>Connect ATG_DBA/ATG_DBA; 3. Get the DDL for the Index before dropping the same SQL> Select dbms_metadata.get_ddl('INDEX','IDX_DCS_PRODUCT','ATG_DBA') from dual; 4. Drop index that is causing the error SQL> drop index IDX_DCS_PRODUCT; Index dropped. 5. Create that index again using the following statement SQL> CREATE INDE
Read more