Path: ...!3.eu.feeder.erje.net!feeder.erje.net!weretis.net!feeder8.news.weretis.net!reader5.news.weretis.net!news.solani.org!.POSTED!not-for-mail
From: Mild Shock <janburse@fastmail.fm>
Newsgroups: sci.logic
Subject: The anchoring problem in a real world philosopher (Re: LLM and
 Prolog, a Marriage in Heaven?)
Date: Thu, 8 Aug 2024 17:00:14 +0200
Message-ID: <v92mhu$vban$2@solani.org>
References: <v67685$6fr5$1@solani.org> <v8ond2$pt0v$2@solani.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Thu, 8 Aug 2024 15:00:14 -0000 (UTC)
Injection-Info: solani.org;
	logging-data="1027415"; mail-complaints-to="abuse@news.solani.org"
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Firefox/91.0 SeaMonkey/2.53.18.2
Cancel-Lock: sha1:EVZ3L1HjatfjTFAwc4x+4pAWcdM=
X-User-ID: eJwFwQkBwDAIA0BL4QutHKDDv4TdhVE46Qx6bOxTZNiIixh5avGh8kJDIblAS40+33tMvQw8PZddFs9a5QcexRRH
In-Reply-To: <v8ond2$pt0v$2@solani.org>
Bytes: 5091
Lines: 128

Hi,

Lets say one milestone in cognitive science,
is the concept of "bounded rationality".
It seems LLMs have some traits that are also

found in humans. For example the anchoring effect
is a psychological phenomenon in which an
individual’s judgements or decisions

are influenced by a reference point or “anchor”
which can be completely irrelevant. Like for example
when discussing Curry Howard isomorphism with

a real world philosopher , one that might
not know Curry Howard isomorphism but

https://en.wikipedia.org/wiki/Anchoring_effect

nevertheless be tempted to hallucinate some nonsense.
One highly cited paper in this respect is Tversky &
Kahneman 1974. R.I.P. Daniel Kahneman,

March 27, 2024. The paper is still cited today:

Artificial Intelligence and Cognitive Biases: A Viewpoint
https://www.cairn.info/revue-journal-of-innovation-economics-2024-2-page-223.htm

Maybe using deeper and/or more careful reasoning,
possibly backed up by Prolog engine, could have
a positive effect? Its very difficult also for a

Prolog engine, since there is a trade-off
between producing no answer at all if the software
agent is too careful, and of producing a wealth

of nonsense otherwise.

Bye

Mild Shock schrieb:
 >
 > Well we all know about this rule:
 >
 > - Never ask a woman about her weight
 >
 > - Never ask a woman about her age
 >
 > There is a similar rule for philosophers:
 >
 > - Never ask a philosopher what is cognitive science
 >
 > - Never ask a philosopher what is formula-as-types
 >
 > Explanation: They like to be the champions of
 > pure form like in this paper below, so they
 > don’t like other disciplines dealing with pure
 > form or even having pure form on the computer.
 >
 > "Pure” logic, ontology, and phenomenology
 > David Woodruff Smith - Revue internationale de philosophie 2003/2
 > 
https://www.cairn.info/revue-internationale-de-philosophie-2003-2-page-21.htm
 >


Mild Shock schrieb:
> There are more and more papers of this sort:
> 
> Reliable Reasoning Beyond Natural Language
> To address this, we propose a neurosymbolic
> approach that prompts LLMs to extract and encode
> all relevant information from a problem statement as
> logical code statements, and then use a logic programming
> language (Prolog) to conduct the iterative computations of
> explicit deductive reasoning.
> [2407.11373] Reliable Reasoning Beyond Natural Language
> 
> The future of Prolog is bright?
> 
> Mild Shock schrieb:
>> Could be a wake-up call this many participants
>> already in the commitee, that the whole logic
>> world was asleep for many years:
>>
>> Non-Classical Logics. Theory and Applications XI,
>> 5-8 September 2024, Lodz (Poland)
>> https://easychair.org/cfp/NCL24
>>
>> Why is Minimal Logic at the core of many things?
>> Because it is the logic of Curry-Howard isomoprhism
>> for symple types:
>>
>> ----------------
>> Γ ∪ { A } ⊢ A
>>
>> Γ ∪ { A } ⊢ B
>> ----------------
>> Γ ⊢ A → B
>>
>> Γ ⊢ A → B           Δ ⊢ A
>> ----------------------------
>> Γ ∪ Δ ⊢ B
>>
>> And funny things can happen, especially when people
>> hallucinate duality or think symmetry is given, for
>> example in newer inventions such as λμ-calculus,
>>
>> but then omg ~~p => p is nevertheless not provable,
>> because they forgot an inference rule. LoL
>>
>> Recommended reading so far:
>>
>> Propositional Logics Related to Heyting’s and Johansson’s
>> February 2008 - Krister Segerberg
>> https://www.researchgate.net/publication/228036664
>>
>> The Logic of Church and Curry
>> Jonathan P. Seldin - 2009
>> https://www.sciencedirect.com/handbook/handbook-of-the-history-of-logic/vol/5/suppl/C 
>>
>>
>> Meanwhile I am going back to my tinkering with my
>> Prolog system, which even provides a more primitive
>> logic than minimal logic, pure Prolog is minimal
>>
>> logic without embedded implication.
>