dennisgorelik: 2020-06-13 in my home office (Default)
~~~
0:42
The system goes online on August 4 1997.
Human decisions are removed from strategic defence.
Skynet begins to learn at a geometric rate.
~~~

That prediction about General Artificial Intelligence systems abilities in 1997 -- looks too over-optimistic in retrospect.
Terminator 2 movie made that prediction in 1991 (6 years prior to 1997).
dennisgorelik: 2020-06-13 in my home office (Default)
Today (and few days ago) I talked with a founder of botson.ai - Dan.
Dan told me how amazing their tech team is in chatbots and jobs matching technology.

When I asked Dan about specific details about what made their job matching algorithms so amazing -- Dan told me that integration of their javascript into our job board is super-easy.

That javascript would tag job seekers who visit our job board with botson.ai cookie.
If job seeker has an open Facebook session - then botson.ai would sign job seeker up for jobs chatbot on Facebook messenger.
Then the chatbot would send jobs that it thinks are relevant to what job seeker is searching for.

I wanted to try that amazing technology before integrating it into our job board and asked Dan to give me a link to signup.
Dan gave me a link, but it did not work. Dan gave me another link. The second link did not work either. The third link worked and I received Facebook message from "Maya - Recruiter".
"Maya" was moderate and only messaged me when I messaged her.
However job matches were not relevant.
I called myself "Monster Botsonai" in a signup form (on the affiliate web site that Dan gave me links me to), so Maya initially called me "Monster", but then corrected itself based on my Facebook profile.

This is my chat with "Maya - Recruiter":
==================
[Maya]
Hi Monster 😊
Following your request for jobs near 32081 here is a selection just for you

[Facebook]
You agreed to have Maya - Recruiter send you updates. Replying will allow Maya - Recruiter to see info you've made public.

[Maya]
Warehouse Associate
We're Hiring Warehouse Associates!. Pay Rate Picking and packing Loading and un...
More Details
Apply Now

[Dennis]
View More Jobs

[Maya]
Pool Technician II
Position Summary:This position is primarily responsible for the daily maintenanc...
More Details
Apply Now

[Dennis]
I look for different jobs

[Maya]
Experienced Auto Hauler
We are a new and growing company that is actively seeking experienced auto haule...
More Details
Apply Now

[Dennis]
[Clicked on "Caregiver jobs"]

[Maya]
Caregiver / Resident Assistant / Personal Care Assistant / CNA
Join a Family, Serve with the Best - Benton House at OakleafAre you a CNA lookin...
More Details
Apply Now

[Dennis]
Can you send me different jobs?

[Maya]
Order Puller - Warehouse
Join the Remedy Family ~ Build Your Career The Remedy Staffing Jacksonville Team...
More Details
Apply Now

[Dennis]
What about C# sql jobs?

[Maya]
Local Landscapers Wanted!
Over 25 million homeowners have trusted HomeAdvisor 1800Contractor to help the...
More Details
Apply Now

[Dennis]
SQL and C#

[Maya]
Hi 😊 I'm an automated HR service and I can only assist you to find more job offers.

I’ll do my best to help you with your search. What job position you are looking for?

[Dennis]
c# sql

[Maya]
Hi 😊 I'm an automated HR service and I can only assist you to find more job offers.

I’ll do my best to help you with your search. What job position you are looking for?

[Dennis]
Software developer

[Maya]
IT Software Developer Senior
EverBank is more than just a bank; it's the evolution of an idea. Traditional fi...
More Details
Apply Now

[Dennis]
C#

[Maya]
Hi 😊 I'm an automated HR service and I can only assist you to find more job offers.

I’ll do my best to help you with your search. What job position you are looking for?

Dennis, was this helpful for you?
Do you want to see more job openings? Just tell me what kind of job position you are looking for 😊

[Dennis]
oncology nurse

[Maya]
Hi 😊 I'm an automated HR service and I can only assist you to find more job offers.

I’ll do my best to help you with your search. What job position you are looking for?

[Dennis]
nurse

[Maya]
Registered Nurse (RN)PRN - Endoscopy
Job ID: 198390Registered Nurse (RN)PRN - EndoscopyJacksonville, FloridaRegular/R...
More Details
Apply Now
==================

So it looks like Botson.ai chatbot recognizes some large job search categories (such as "software developer" and "nurse"), but does not recognize more specific skills (C#, SQL, oncology).

My conclusion is that the intelligence behind the chatbot technology is not ready yet, but the idea to use chatbot for job search - looks interesting.
dennisgorelik: 2020-06-13 in my home office (Default)
Meet Matt:
1) Advanced Artificial Intelligence researcher.
2) Created the best data compression algorithm for 7-Zip (and made a small fortune in process).
3) Smart.
4) Open.
5) Honest.
6) In a great physical shape:

- Does Yoga:



- Runs marathons and super-marathons:


7) Social and nice to people.




You may assume that Matt is so good that he has no problem with having a satisfying love life.

But that is not the case.

Matt is 61 and so far he had two love relationships:
1) At 37 [sic!] Matt met and eventually married Joan.
Joan was much older than Matt, was verbally abusive and refused sex for the last several years of their relationship - until her death from a stroke.



2) For the last 5 months Matt dated Karen and describes that relationship as the happiest of my life


Karen left Matt (and Florida) for her daughter in Maine.

==============
https://www.facebook.com/mattmahoneyfl/posts/10212741387674099?comment_id=10212751897296833
I was socially inept until my 20's and single until I met Joan at 37. We had an emotionally abusive marriage for 22 years before she died 3 years ago. We fought and yelled all the time until the neighbors would call the police. She was depressed, held resentments, and was never happy, which is a common problem among sober alcoholics from alcoholic families. She also had PTSD and nightmares and became extremely racist after a black man broke into her home and raped her years before we met. She hated people, especially men, especially all my friends and family, and most especially me. When my mother would send me a birthday card, she would get the mail first, write vile obscenities on the envelope and send it back before I saw it. She would walk up to fat people, complete strangers, and tell them they were fat and they should die. Maybe it was her insecurity about her fitness. She was very goal oriented, got a Ph.D. in psychology, set perimeter bicycling world records (24 countries), indoor rowing records (33M meters), and completed Ironman 3 times. I saw her 5K times go from 21 minutes to over an hour as her health deteriorated before she died at 76. Maybe she hoped that people would like her for her accomplishments but it doesn't work that way. She refused to let me have time with friends. She would get enraged if I wanted to go to a fun run or if I didn't leave a race immediately after she got her award. She refused sex with me for the last several years of her life. When she died I should have grieved, but it felt more like I was let out of prison.

I don't know why I didn't leave her. I was always mystified when I would read about women that stayed with abusive men, but now I realize it works both ways. I don't like being single, but dating scares the shit out of me. On the list of things I want to do, dating ranks between hiring a lawyer and cancer surgery. I don't know why women think it is easy for men to meet women to date.

But I tried anyway. Karen was the first to show any interest in me after being single for over 2 years. Our time together was wonderful. She likes my friends and I like hers. If I ever did anything to piss her off, she would let me know right away and we would resolve it, instead of holding a resentment and bringing it up years later. I can trust her completely, something that Joan never did for me because so many other people had let her down in her drinking days. I still love Karen and I keep hoping she will change her mind and come back. But if she doesn't I will have to accept it and move on.
==============

That story is quite impressive and made me think.
Here is my analysis:
1) It is not enough to be a perfect man in order to have a good love relationship.
You have to actually find a good match.
2) In order to find a match you have to put an effort into matchmaking:
- Think about how and where to find the right person.
- Put time and effort into actual search.
- Convince that person to stay with you.
- Train yourself to handle rejection (not everyone who you consider a good match would consider yourself a good match, but until you ask - you may not know).
3) That lesson can be applied not only to romantic relationships, but to all other relationships: work, (jobs and business), hobbies and friendship.

What do you think: is ability to make new relationships an important skill?
dennisgorelik: 2020-06-13 in my home office (Default)
Business context
For years I wanted to collect new jobs from all over internet in order to send appealing job alert emails to candidates that created a profile on postjobfree.com
So, finally, I decided to create a web crawler for that.
However, unlike Google, I do not want to crawl billions of pages (too expensive). Several million pages should be good enough for the first working prototype.
The question is - how to determine automatically what pages to crawl and what to ignore?
That's why our web crawler is combined with self-learning neural network.

Data structure
We represent every page as a record in PageNeuron table (PageNeuronId int, Url varchar(500), …, PageRank real, ...)
We represent links from page to page in LinkAxon table (..., FromPageNeuronId int, ToPageNeuronId int, …)

PageRank calculations
PageRank is inspired by classic Google PageRank, however we calculate it differently.
Instead of calculating probability of visitor click, our NeuralRewardDistribution process distributes PageRank from every PageNeuron record to every connected record (in both directions).
With every “reward distribution” iteration NeuralRewardDistribution process distributes about 10% of PageRank to other pages (that amount is split between all reward destination PageNeuron records proportionally to LinkAxon weights).
Then, in order to prevent self-excitation of the system, NeuralRewardDistribution applies "forgetting" by reducing PageRank of original page by 10%.

Setting goals
When NeuralPageEvaluator parses crawled pages - it tries to detect words and patterns we need.
Every time NeuralPageEvaluator finds something useful, it adds reward in form of extra PageRank for the responsible PageNeuron record. For example, we reward:
- 1 PageRank point for such words as "job", "jobs", "career", "hr".
- 10 PageRank points for such words as "hrms", "taleo", "jobvite", "icims".
- 1000 PageRank points when parser discovers link to a new XML job feed in the content of PageNeuron record.
- 20 PageRank points when parser discovers link to an XML job feed that we already discovered in the past.

What to crawl
NeuralPageProcessor processes already crawled pages (PageNeuron) by passing them to NeuralPageEvaluator.
NeuralPageEvaluator returns collection of outgoing links from that parsed page.
If extracted outgoing link is new, then NeuralPageProcessor creates new PageNeuron record for it. For initial PageRank it uses (10% of source PageNeuron record PageRank * link share or that new URL among all other URLs that source PageNeuron record points to).
NeuralCrawler crawls new PageNeuron records with highest PageRank.

Cleanup
NeuralRewardDistribution deletes PageNeuron records (and all corresponding LinkAxon records) of PageRank is too low.
Current "delete threshold" is at PageRank = 0.01 which deletes about half of ~3 million PageNeuron records we already created.

Profile

dennisgorelik: 2020-06-13 in my home office (Default)
Dennis Gorelik

June 2025

S M T W T F S
1234 567
891011 12 13 14
15161718192021
22232425262728
2930     

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jul. 8th, 2025 07:14 am
Powered by Dreamwidth Studios