Hi Friends,

Even as I launch this today ( my 80th Birthday ), I realize that there is yet so much to say and do. There is just no time to look back, no time to wonder,"Will anyone read these pages?"

With regards,
Hemen Parekh
27 June 2013

Now as I approach my 90th birthday ( 27 June 2023 ) , I invite you to visit my Digital Avatar ( www.hemenparekh.ai ) – and continue chatting with me , even when I am no more here physically

Friday, 11 June 2021

EXPERT SYSTEMS - RICHARD FORSYTH



====================================================================================

When I read any book, I scribble my comments / notes in the margins


These reflect my views / opinions about what the author is saying – including my disagreement

Often, my comments are in the nature of telling myself :

Hey ! We should try out this idea in our own business ( Head-hunting / Online Recruitment )

Following are my comments re :

EXPERT SYSTEMS - RICHARD FORSYTH      

                                                                                                          

=====================================================================================

Page no. 3

·         I read his “Human use of human Beings,” around 1957/58. Wiener was called “father of cybernetics”- in fact, he coined the word “Cybernetics.” (30-08-2002).

Page no. 6

·         Is this like our saying “If such & such keywords appear in a resume, THEN it may belong to such – Industry or Function? (30-08-02).

Page no. 7

·         It is another matter that DEC went bankrupt & was taken over!

·         I believe the “knowledge contained in our 65000 resumes, is good enough to develop an expert system (ARDISARGIS)

·         We started work on ARDIS/ARGS in 1996!

·         But taken-up seriously, only 3 months ago. (30-08-02).

Page no. 8

·         “Keywords” are nothing but “descriptions of resumes.

Page no. 9

·         This was written 13 years ago. May be, by now, there “problems” have been solved.

Page no. 10

·         Back-end

·         Front –end

Page no. 11

·         Inder/Anjaria/our consultants

·         I believe ISYS manual speaks of “context tree” so does oracle context cartridge (Themes)

Page no. 14

·         As in our case

Page no. 15

·         In our case

·         “Hypo the sized outcome” could be a resume

·         Getting shortlisted by 3p

·         “          “          “          Client

·         Or

·         A candidate getting “appointed (after interview)

·         In our case

·         The “presence of the evidence” could be presence of certain “keywords” in a given resume (the house) or certain “Edu. Qualification” or certain “age” or certain “Exp (yrs.)” or certain “current Employer”etc. etc.

Page no. 16

·         In our case, these several “pieces of evidence “could be

·         Keywords

·         Age

·         Exp.

·         Edu Quali.

·         Current Industry back grad

·         Current Function background

·         Current Desig. Level

·         Current salary

·         Current Employer etc. etc.

·         We could “establish ODDS (for each piece of evidence) and then apply “sequentially” to figure out the “probability/ODDS” of that particular resume getting shortlisted/getting selected

·         We have to examine (atatistracally) resumes of all candidates shortlisted during last 13 yrs. to calculate “ODDS.” (30-08-02).

 

Page no. 18

·         “Automating the process of Knowledge Acquisition? We could do this (automating) by getting the jobseeker to select/fill in keywords online, in the web from. (31-08-02).

 

Page no.19

·         I suppose this has become a reality during the last 13 years since writing of this book.

·         The “Decision support”

·         That our consultants need is:

·         “From amongst thousands of resumes in our databank, which “few” should be sent to client? Can software locate those few, automatically, which have “excellent probability” of getting shortlisted/selected?

·         Our consultants, today, spend a lot of time in doing just this, manually- which we need to auto mate.

 

Page no. 20

·         These (few) resumes are “GOOD” for this “vacancy.”

 

Page no. 22

·         According to me, this “notation” is:

·         All human thoughts/speech & Action, are directed towards either increasing the happiness (of that person) or towards decreasing the pain, by choosing from amongst available thought/spoken words/Actions” this notation describes all aspects   of human-race.

·         This ability to choose the most appropriate option (at that point of time), make a human being “intelligent.” (31-08-02).

 

Page no. 23

·         There are millions of “words” in English Language-used by Authors in Books & poets in songs & Lawyers in documents but the words of interest to us are these used by jobseekers in Resumes & Recruiters in job-advts.  This is our area expertise

·         Probabilities of 10000 Keywords occurrence amongst “past successful” candidates.

·         See remarks at the bottom of P=19 for “OUR” problem- description.

 

Page no. 25

·         “RESUMIX” (resume management software) claims to contain 100,000 “rules.”

 

Page no. 26

·         Our expertise in “match making” of jobseekers & “vacancies” of Recruiters.

·         Our business does fall in such a “specialist” category.

·         Person who have spent is years reading resumes/deciding their “suitability & interviewing candidates.

 

Page no. 27

·         See remarks at bottom of p=19

·         In der & May be some of our own consultants.

·         Agree. We do not expect “Expert system” to conduct interviews!

·         Our consultants do spend 2/3 hours daily in reading/shortlisting resumes.

·         Abhi, Rajeev

·         We want a “Decision support system” to assist our consultants so they can spend more time in interviews type of assessment.

·         If, during last 13 years, we have placed 500 executives, then we/client must have short-listed 5000 resumes. These are enough “Test cases.

·         Must be several hundred by 2002.

 

Page no. 28

·         In last 13 years, this has grown may be 50 times! 50 that cannot be a limitation

·         Now we have 4th generation. Now we have PENTIUM 4.

·         I had perceived this as far back as 1996.

·         Now (in 2002), expert systems, have become an “Essential” to survival of all organizations. We can ignore it at our peril.

 

Page no. 29

·         We can become VICTORS-or the VICTIMS: choice is ours.

·         I am sure by 2002, we must have many “MATURE” expert system “keywords” “shells,” commercially available in the market.

·         New available for $40. (3000).

·         Abhi, Rajeev, Deepa+ Inder IIT (Pawai),

·         We don’t need but we could talk to TIFR or NCST professors of AI/Expert system for “guidance”(31-08-02).

 

Page no. 30

·         Ask NCST (juhu scheme) if they can train us.

·         May be we could send an email to Mr. FORSYTH himself, to seek his guidance. We will need to explicitly state- our problem – solution which we seek from the Export system & ask him which commercially available “shell” does he recommend.

·         Email: Richard. Forsyth@ uwe.ac.uk. (31-08-02).

 

Page no. 31

·         Which it actually did

 

 

Page no. 32

·         WWW.scientio.com

·         How many does this Directory list in 2002?

·         Google still shows CRI-1986 as the latest! But “Expert system” in Google returned 299000 links!

·         I took a course in x- ray crystallography at KU in 1958. (31-08-02).

 

Page no. 33

·         When developed, our system would fall under this category.

·         Most certainly we should integrate the two.

 

Page no. 35

·         The resumes shortlisted by our proposed “Expert system” (resume having highest probability of getting shortlisted), must be manually “examined” –and assigned “weightage,” by our consultants & these “weightages” fed back into the system.

 

Page no. 37

·         This may have happened by now. By now, quite cheap

·         Abhi, Rajeev

·         I believe our system will be simple “rule-based”- although, there may be a lot of “processing” involved, in “sequential” computation of profanities for Keywords related to :

·         Industry/function/Designation-level/Age/Exp/Edu.Quali./Attitudes/attributes/skills/Knowledge/salary/current Employer/current posting location/family. Etc. 

 

Page no. 38

·         “RESUMIX” claims to have 100,000 rules.

 

Page no. 39

·         Abhi, Rajeev

·         In my notes on ARDIS/ ARGIS, see separate notes on “logic for…. Here, I have listed the under lying rules. (31-08-02).

 

Page no. 40

·         Expert Knowledge (-and consequently the rules) contained in RESUMIX have relevance to USA jobseekers –and their “style” of resume preparation these (rules) may not apply in Indian context.

·         Page no. 41

·         We are trying to establish the “relationship” between:

·         Probability of occurrence of a given “keyword” in a given resume

·         Probability of such a resume getting “shortlisted.” 

·         Eg: if we have lost the resumes of 5000 candidates who got shortlisted during last 13 years!

·         Only “Age” & “Exp (yrs)” are dependent in our case.

Page no. 42

·         Exp (yrs) can never be > Age (yrs)

·         So, we will need to prepare a comprehensive list of inconsistences with respect to a resume,  eg: as shown above. This will be an ongoing take for Inder. (01-09-02).

Page no. 43

·         Abhi, Rajeev

·         We should ask both (the Expert system& the experts) to independently short list resumes & compare.

·         We have to experiment with building of an expert system which would “ test/validate” the assumption:

·         If certain (which?) keywords or search parameters are found in a resume, it has a higher probability of getting/selected.

Page no. 44

·         eg: system shortlisting  a “sales” executive against a “production” vacancy!

·         What/which “cause” could have produced what/which “Effect/Result.”

Page no. 45

·         “Human use of Heman beings” by Norbert wiener.

·         In our case, the expert system should relieve our consultants to do more “In Telligent” work of assessing candidates thru personal  interviewing.

Page no. 47

·         Eg: Entering email resumes in “structured” database of module I

·         Reconstituting a resume (converted bio data) thru ARGIS, automatically

·         For these tasks, we should not need human beings at all.

·         Read “what will be” by Michael Dertouzo (MIT lab) 1997

Page no. 48

·         Even when our own Expert system “shortlists” the resumes (based on perceived high probability of appointment), our consultants would still need to go thru their resumes before sending to clients. They would need to “interpret.”

·         No wonder Google retums 299000 links when you search for “Expert systems”! Forsyth had foresight!

·         This is Inder`s main task.

·         This is ME!

·         Read all of my notes, written over last 13 years. (01-09-02).

Page no. 49

·         My notes

·         Inder, Abhi, Rajeev

Page no. 50

·         Our future/new consultants need to be taken thru OES, step-by step thru the entire process –thru SIMULATION fake search assignments   

·         Our “Task-Area” is quite obvious but may not be simple viz: we must find the “right” candidates for our clients, in shortest possible time.

·         In 13 years, since this book was written, “mobile computing” has made enormous strides. Also internet arrived in a big way in 1995. By march 2004,I envisage, our consultants, carrying their laptops or even smaller mobile computers & search our MAROL data base for suitable candidates (of course using Expert system), sitting across client`s table.

Page no. 51

·         We can repeat this even in 2002!

·         Our consultants’ productivity

·         These are our objectives.

·         To automate as many “business processes” as possible.

Page no. 57

·         Data bases & knowledge bases.

Page no. 58

·         Abhi, Rajeev

·         Resumes are “data” but when arranged as a “shortlist” they become “information” because a “shortlist” is always in relation to our “search – assignment”! it is that search-assignment that lends “meaning” to a set of resumes.

 

Page no. 59

·         Are “Resumes,” Knowledge about “people”? & their “Achievements?

Page no.60

·         But is a human, a part & parcel of nature? Human did not create the human? Our VEDAS say that the entire UNIVERSE is contained in an ATOM may be they meant that an entire UNIVERSE can arise from an atom!

·         Experience *Age

·         Are 204 industries – names&110

·         “Function-name” granular enough? Can we differentiate well? (08-09-02).

Page no. 61

·         Read “aims of Education” by A. N.  WHITEHEAD

·         Inference process of Drawing/Reading “conclusion “based on Knowledge

Page no. 62

·         Calculating probabilities of occurrence of keywords & then comparing with keywords contained in “Resume” of successful candidates

Page no. 64

·         If a resume R contains keywords

·         a,     b,     c  

·         And if resumes of all past “successful

·         Candidates also contain keywords a, b, c

·         Then the chances are that

·         Contusion:-

·         Resume R will also be such an

·         “Successful”.

·         Our expert system will be such an “Automatic Theorem proving system”

·         Where   “Inference-rules” will have to be fist figured-out/established, from large volumes of past co-relations   between “keywords” & “successes.”

Page no. 65

·         Abhi, Rajeev

·         “Successes” can be defined in a variety of ways, including :

·         Shortlisting

·         Interviewing

·         Appointing etc. et

Page no. 66

·         Abhi, Rajeev

Page no. 67

·         In our case too, we are trying to “Interpret/diagnose” the “systems’ (in our case-the keywords)

·         Contained in any given resume (patient?)& then “predict” what are  its chances (i.e.(=getting cured)

·         i.e. getting shortlisted or

·         Getting interviewed or

·         Getting appointed

·         (08-09-02).

Page no. 68

·         These resumes can be further sub-divided according to

·         Industry

·         Function

·         Design-level etc. to

·         Reduce population size of keywords.

·         For us, there are as many “rules” as there are “keywords”

·         In the resumes of past “successful” candidates with the “frequency of occurrence” of each such keywords of (in, say, 7500 successful resume), deciding its “weightage ”while applying the rule.

Page no. 69

·         Our initial assumption – Resumes of “successful (past) candidates” always contain a, b, c, & d,

·         Process – Find all other resumes which contain a, b, c,& d.

·         Concision- These should “succeed” too.

·         Our Good –find all resumes which have high resume probability of “success.

·         System should automatically, keep adding to the database all the actually  successful candidates as each search assignment gest over.

Page no. 70

·         In 1957, this was part of “operations Research” course at uni: of Kansas.

·         With huge number, crunching computers, computational costs are not an important consideration anymore.

Page no. 71

·         Cut-finger & Pain follows (consequence) somewhat similar to our situation of resumes & keywords.

·         (10-09-02).

 

Page no. 72

·         We plotting frequency of occurrence of keywords in specific past resume to generalist our observations

·         We can consultants a take like this from our 65000 resumes & then try to develop “algonthms.

·         Abhi, Rajeev

·         In above table the last column (Jolo or Actual Designation can also be substituted by & a new set of “algorithms “will emerge.

·         Industry

·         Function

·         Edu.Quali

 

Page no. 74

·         Abhi, Rajeev

·         Our resumes also “leave out” a hell of a lot of “variable”! A resume is like a jigsaw puzzle with a lot of missing pieces! We are basing on statistical forecasting viz: frequency of occurrence of certain Keywords & (attaching) probability values.

Page no. 75

·         This statement must be even more true today -13 years since it was fist written

Page no.76

·         Just imagine, if we too can locate & deliver to our client, candidate that he need! In the first place just those resumes, which he likely to value appreciate.

·         I am sure by now superior hardware certainly has, as has “conventional Tools” of database management.

Page no. 77

·         Perhaps what was “specialized” hardware in 1989, must have become quite common today in 2002 and fairly cheap too.

Page no. 81

·         We must figure-out (-and write down), “logic/rules” our consultants use (even subconsciously), while selecting/rejecting a resume (as being “suitable”) for a client need. Expert system must mimic a human expert.

·         We are basing ourselves (i. e. our proposed expert system) on this “Type” (see “patterns “in  “Digital Biology”)

 

Page no. 82

·         This is what we are after.

Page no. 87

·         Industry/Function etc.

·         Keywords?

·         Probability that this keyword (symptom) will be observed/found, given that the resume (patient) belongs to XYZ industry (lines)

Page no, 88

·         Random person = any given “incoming” email resume Influenza =”Automobile” Industry

·         Based on our past population? =”Auto” Resume All resumes probability

·         If symptoms =keywords, which symptoms (keywords) have appeared in “Auto” resumes or which keywords have never appeared in “Auto” resumes OR appeared with low frequency?

Page no. 89

·         As demon started in “Digital biology”

·         With addition of all keywords (including new keywords not previously recorded as “keywords”)  

·         From each new, incoming resume, the “prior probabilities” keep changing. (13-09-02).

Page no. 90

·         In this “Resume,” it belongs to a particular Industry

·         What keywords may can be expected

·         What keywords may not be expected We can also, resume the “Reasoning” viz:

·         What “Industry” (or Function) might a given incoming resume belong to, if,

·         Certain keywords are present and

·         Certain keywords are absent? The “result/answer” provided by Expert system can then be tested/ verified with what jobseeker has himself “clicked.”

Page no. 91

·         Re-iteration so each new incoming resume would change the “prior probability” (again & again) for each

·         Industry

·         Function

·         Designation

·         Edu. Quli

·         Exp. etc. etc.

Page no. 92

·         With 65000 resume (i.e. patients) & 13000 keywords (systems) we could get a fairly accurate “estimate” of “prior probabilities” this will keep improving/converging as resume as resume database & keywords database keeps growing especially if we can succeed in down lodging thousands (or job-davits) from Naukari/monster jobs ahead etc. etc.

·         Eg: Birthdate as well as Age.

·         This is good enough for us. (13-09-02).

Page no. 93

·         Eg: In Indian Resumes, keywords keyword “Birthdate” would have the probability 0.99999. of course most  such keywords are of no interest to us!

Page no. 95

·         For our purpose, “keywords” are all “items of evidence.” If each & every “keyword” found in an (incoming resume, corresponds to our “hypothesis (viz: keywords, A. & B. & C are always present in resumes belonging to “Auto industry) then we obtain max. ”Posterior probability”

·         So, if our knowledge base (not only of keywords, but phrases/names of current & past/sentences/employers/ posting cities etc.) is VERY WIDE & VERY DEEP, we would be able to formulate more accurate hypothesis & obtain higher posterior probability. (14-09-02).

Page no. 96

·         So, the key issue is to write down a “set of Hypothesis” 

Page no. 97

·         Let us say, keyword “Derivative” may have a very low frequency of occurrence in 65000 resumes (of all Industries put together)  but, it could” be a very important keyword for the “financial service” Industry.

Page no. 98

·         Eg: certain key words are often associated (found in) with certain “Industries “or certain “Function” (Domain Keywords).

·         Page no. 99

·         With each incoming resume, the probability of each keyword (in the keyword data base) will keep changing.

·         Eg: Does “Edu Quali.” have any role/effect/weightage in the selection of a candidate?

·         Eg; what is the max .Age (or Min, Exp) at which corporates will appoint a

·         Manager

·         General Manager

·         Vice – President

·         (23-09-02).

Page no. 100

·         Line of Best Fit?

Page no. 103

·         Hypo theses

·         Evidence

·         Say, every 20th incoming resume

·         Based on frequency

·         Distribution in existing 65000 resumes.

·         Belongs to

·         XYZ Industry

·         ABC Function

Page no. 104

·         Evidence: If an incoming resume belongs to “Auto” industry it would contain keywords “car”/”Automobile/or Edu. Quali =Diploma in Auto –Engineering

·         Eg: probability of selection of an executive is O (zero)if he is above 65 yrs. of age!

·         One could assign “probability of getting appointed” for each “Age” or for each “years of Experience” or each “Edu. Level etc. etc.

·         This is very similar to asking: “will this executive (incoming resume) NOT get shortlisted? OR NOT get appointed?”

Page no. 105

·         Obviously

·         A person (incoming resume) having no experience (fresh graduate) will NOT get shortlisted for the position MANAGER (zero probability)

·         A person (incoming resume) having NO  graduate degree, will NOT get shortlisted for the position of MANAGER (zero probability)

·         A person (incoming resume) with less than 5 yrs. of experience will NOT get shortlisted for the position of GENERAL MANAGER (zero probability) but , will get shortlisted for the position of SUPERVISOR(o.9 probability) etc. etc. (23-09-02).

Page no. 106

·         So, we need to build-up a database of who (executive) got shortlisted/appointed by whom (client) & when & why (to lest of our knowledge) over the last 13 years & how much his lock ground matched “client requirements”& keywords in resumes.

 

Page no. 107

·         How “good” or “bad” is a given resume for a given client need?

·         Co-relating 1. Search- parameters used, with 2. Search-results (resume/keywords in resumes) for each & each very, online/offline Resume-search.

Page no. 108

·         Now 37 yrs. Even in washing machines

·         Is this like saying,” resume A belongs to Engineering Industry, to some degree and also to Automobile Industry to “some degree?

·         Some as our rating one resume as “Good” & another as “Bad” (of course in relation to clients stated needs).

Page no. 110

·         Eg; set A – Resumes belonging to Eng. Industry & set B – Auto

·         A resume can belong to both the sets with different degree of membership”

Page no. 114

·         In our case we have to make a “discrete decision

·         Viz: shall I shortlist & forward this resume to client, or net?

Page no. 115

·         A given keyword is present in resume, and then treat the resume as belonging to “Eng” industry

·         OR (better) a given “ combination of keywords being present or about in a resume, shall decide whether that resume should be classified/under (sey) “Eng” Industry.

·         For us “real world experience “ =several sets of “keywords “discovered in 65000 resumes which are already categorized (by human experts) as belonging to industry (or function) A or B or C.

Page no. 118

·         Eg: A given resume is a

-          Very close match

-          Close match

-          Fairly close match

-          Not a very close match

·         (With client’s regalements)

Page no. 119

·         I am sure, by now, many more must be commercially available.

 

Page no. 127

·         Who knows, what ‘new/different “keywords would appear (-and what will disappear) in a given type of resumes, over next 5/10 years?

·         We may think of our corporate database in these & call it “corporate Knowledge base.

Page no. 129

·         One could replace

·         “colour – teeth=

·         “Designation level”

·         =MD

·         OEO

·         President

·         Gen-mgr. etc.

·         Worth checking-now those 14 yrs have elapsed.

Page no. 130

·         Have they?

Page no. 132

·         Is this somewhat like saying “this (given) resume belong to Industry

Page no. 133

·         Which is what the resumes are made up of (-sentences). See my notes on ARDIS & ARGIS

Page no. 136

·         This was written 13 yrs ago. In last few months, scientists have implanted simple microprocessors in human body& succeeded in “integrating” these (chips) into human nervous system!

Page no. 139

·         There is no doubt pronation must have made tremendous progress in last 13 years!

·         We now have “Gigabyte”(1000 times bigger)

Page no. 140

·         Every major nation has a “weather forecast satellite these days.

Page no. 143

·         In ARDIS I have talked about character recognition/word recognition/phrase sentence recognition

Page no. 178

·         We are thinking in terms of past resum,es which have been “shortlisted” I have already written some rules.

·         This must have happened in last 13 years!

Page no. 179

·         We must question our consultants as to what logic/ rules do they use/ apply shortlist or rather select from shortlisted candidates

Page no. 180

·         This list must have vastly grown by now.

Page no. 186

·         I have written some rules but many more need to be written.

Page no. 191

·         Like weightages in neural networks?

·         To find “real evidence” take resumes (keywords) “interview Assessment sheets” of “successful” candidates & find correlation.

·         Eg: Rules “Designation –level could conflict with rules

·         “Experience (years) or “Age (Min)”

Page no. 192

·         Basic degree in commas (B.com)

·         “Rule sets on

·         Age (Max)

·         Exp (Min)

·         Edu.Quali

·         Function

·         Designation etc.

·         If our corporate client can explicitly state “weightages for each of the alove it would simplify the design of expert system mr. Nagle had actually built this (weightages) factor in our HH3P search engine way back in 1992.

Page no. 197

·         Quacks Vs Doctors?

·         Quacks sometime do “cure” a disease, but we would never “Know” why or how!

Page no. 198

·         We have databases of

·         Successful (or “failed”) candidates, and

·         Their “resumes” & “Assessment sheets” and

·         Keywords contained in these resumes

Page no. 199

·         We must have data on past 500 successful & 5000 “failure” candidates in our database

Page no. 219

·         This has happened by now

Page no. 222

·         Abhi

·         This is exactly what our knowledge Tools do:

·         Highlighter

·         Eliminator

·         Refiner

·         Match maker

·         Educator

·         Compiler

·         Member- segregators etc.

Page no. 223

·         We are one.

Page no. 224

·         Resume & job advt. databases on jobsites

Page no. 227

·         This was written in 1989!

·         e- Commerce defined 13 yrs. ago.

·         “ SiSA” on jobstreat.com has elements

·         Of such a self- serving system.

 

Page no. 231

·         Our knowledge Tools, do this.

Page no. 232

·         Statistical analysis of thousand of job advts of last 5 years, could help us extrapolate next 10 year`s trend.

Page no. 236

·         We are planning a direct link-from HR mgr. to our OES-(which is our factory floor)

·         We will still need “consultants” but only to interact (talk) with clients & candidates not to fill in IN PUT screens of OES!

Page no. 237

·         Obviously Author had a vision of Interest

·         Which 1 could envision 2 years earlier “QUO VADIS”

  Page no. 239

Norbert wiener had predicted this in 1958.             


No comments:

Post a Comment