EFTA00767968.pdf

From: roger schank To: Jeffrey Epstein <jeevacation@gmail.com> Subject: Re: beating a dead horse Date: Sat, 31 Oct 2009 11:47:55 +0000 spoken like someone who has never advised PhD students; it is a beginning, not an end; because you can think faster and deeper than he does proves nothing but there is a point here you are not getting which is the how; you have to set up the right question to find answers; if you want a smart SI system that detects errors on a ship you need to work out the details of the how which is what he is doing and my lower level of acceptance of intelligence that you cite comes not from softball, since I dont talk to those people, (who could?) but from graduate students I had over 35 years who are nowhere near as smart as dimitri roger schank On Oct 31, 2009, at 7:42 AM, Jeffrey Epstein wrote: You have lost your mind„ , this is total bullshit.. goals are not one dimensional. actor goals specifically, the captain might want money, promotion, bribe, sleep . respect, .. they all will interfere, with the enterprise goal.. with randomness, and varying amplitudes.. what the military does is say , if you have any other goal than mine , you're dead. so have none. Roger this type of blather, is not worth you spending two seconds on. you will drill down and find more to drill.. I'm sorry, for you to think that his description of how peoples motivation can influence outcome is borderline retarded. I think your softball conversations have reduced your level of : smart " criteria , to an absolute minimum.. This could have been written by a first year schvartza at howard university and received a Bminus On Sat, Oct 31, 2009 at 7:32 AM, roger schank > wrote: this is dimitri working on an idea he needs for shipping that can be inferred from ideas of mine in Dynamic Memory my comments, if he were a PhD student of mine here, would been that this idea would work well as a PhD thesis and he should continue on this road, but it would a good idea to take a course in how to write clearly this is what smart looks like to me; it is not what brilliant looks like, but how many brilliant people are there really? I will settle for smart roger Begin forwarded message: EFTA00767968 From: Dimitris Lyras Date: October 30, 2009 9:52:35AM EDT CA To: "roger Subject: goa ca cu us How does this sound chief? Long time no hear. How was your home alone week? Any hanky panky....?) How do actor goals affect process steps and process goal conflicts? Actors affect the success of the subordinate steps they control in a process. If their goal in completing the step are not aligned to the enterprise goal of that process, then the goals of individuals with affect on the steps in the process may alter their behaviour towards completing the step and thus alter the probability a typical step hazard occurring and effecting process goals.. So there is an interaction between process goals and individual actor goals and step problems. The goals are in turn influenced by values of enterprise and individuals. So we need the relationship between actor gaols and probability of step execution problems when addressing processes where the standard process goals do not align with those of the actors. This is similar to actor skills, we expect that actor skills affect probability of failure without the need for a conflict with the process goal. So we need a role relationship to process steps. The role has values and the role has skills. Both skills and values affect the probability of gaol step failure. So process step execution needs probabilities for customary problems and these probabilities are affected by actors. Example; A truck driver wants to reach his destination quickly. His perception of safety and its importance is low. The enterprise for which he works puts safety first. The goals are in conflict. The driver influences the speed of travel while the enterprise is not there to do so. So the risk of fast reckless driving in the driving step of the road haulage process is exacerbated. Its proximity to the goal of road haulage does not change but the probability of failure does. But is the gaol conflict at the driving process level or the haulage process level. In other words is the goal conflict at the step or the main process level. It is actually at both in this case. Let's take another example; the cancelling date agreement in a sale and purchase contract; The agreement process has actors influencing each term. The actors are in conflict when prices change in the market. Their actions influence the cancelling date in opposite ways. The goals of the bigger process influence the goals of the step. So the main process goals are in conflict or not aligned and so the step goals follow inheriting the misaligned goal hierarchy; The actual actors on each step are the most influential and so the problems in the process may be exacerbated from a probability standpoint. But which ones? Watch-keeping requires attention and if the ship is unstable the deck watch will be less attentive. But other watch-keeping hazards will not be more probable, for example the windscreen will not be any less clear or the probability of windshield wiper operation. The probability only applies to the steps influenced by the actors goal conflict. EFTA00767969 We already have probability calculated by the system or entered by specialists applying to problems (stories). Now we need probabilities to be adjusted by actor goal conflicts and actor skills. Actor values and goals need to be compared to the process goals in order to ascertain the conflict. If the conflict exists then one of the process goals or conditional goals is more vulnerable. This vulnerability is quantified by the degree of actor influence on a step and the probability of increase of a predetermined potential step failure This occurs at the analysis of each process and its process steps. The probability of step failure under typical failure events is done separately. Then each one is given a degree of potential influence from actor skills and actor values. Conflicts of interest are at the core of finding trouble areas in a management system. The processes involving them need allot of attention. The result is that certain otherwise improbable problems become far more probable. In risk management this is a key issue. In retrieval of similar cases it applies if the step failure and generic step failure are the same type of expectation failure. In these cases the similar cases are considered normally via same process and then generic process and goal proximity and then the gaol conflicts are investigated for further matching. How do step failures influence gaol hierarchy: When steps fail processes are reorganised as a plan. So planning requires a new gaol hierarchy based on the circumstances of failed steps etc. Using Case retrieval for planning: I. Make a goal hierarchy for the plan 2. Make preliminary plan from existing processes. In a well know domain with mature processes and good cause and effect experience this is not complex. In domains where cause and effect are hard to clearly identify (as in economics or in new industries) or processes are not mature, we need to use processes that are more concrete as processes to put in the plan. In an immature field existing processes will not achieve the goals in that domain so there will be a need to modify those processes such that they achieve the goals. Then we will need to take the new changed steps in the modified processes and ascertain their interactivity with other steps in the process and their goal proximity to the plan goals. Then we will need to figure out their generic process components. This is time consuming but could be appropriate. 3. Use existing process relationships to retrieve stories 4. In the interaction between processes for the plan, note the goal proximity 5. Find parallel generic processes to the plan 6. Look at process steps and recalculate their proximity to the plan goals EFTA00767970  7. Find generic process steps within these processes that can cause goal failure. 8. Find same generic process steps in another domain which affect similar generic processes with similar proximity to goal. 9. Look for actor goal conflicts as retrieved from processes used to make the plan and applying the users common sense. The information contained in this communication is confidential, may be attorney-client privileged, may constitute inside information, and is intended only for the use of the addressee. It is the property of Jeffrey Epstein Unauthorized use, disclosure or copying of this communication or any part thereof is strictly prohibited and may be unlawful. If you have received this communication in error, please notify us immediately by return e-mail or by e-mail to jeevacation@gmail.com, and destroy this communication and all copies thereof, including all attachments. EFTA00767971