Can computers compose artistic problems? (1)

by GM Pál Benkö
6/14/2016 – Some time ago Dr Azlan Iqbal presented a program, Chesthetica, that was composing chess problems. We published ten examples of three-movers by the machine. Now a leading expert in the subject, Pal Benko, who is one of the finest problem composers in the world, tells us what he thinks about the quality of the computer compositions – and also what are the criteria that make a chess problem valuable.

Can computers compose artistic chess problems? Part one

By Pal Benkö

An article on the ChessBase news page of 11/7/2014 was entitled "A machine that composes chess problems" raised my interest. I quote from its introduction.

Chess problems are an art – positions and solutions, pleasing to the mind and satisfying high aesthetic standards. Only humans can compose real chess problems; computers will never understand true beauty. Really? Dr Azlan Iqbal, an expert on automatic aesthetic evaluation, imbued his software with enough creativity to generate problems indefinitely. The results are quite startling.

A new artificial intelligence approach called the DSNS (digital synaptic neural substrate) was developed that enables a computer to become ‘inspired’ by the different types of objects, fed into it, such as photographs of people, painting masterpieces, classical music, chess tournament games and other chess problems.

In his article Dr Mohammed Azlan Iqbal presented ten selected work composed by the machine:

Chess problems created by Chesthetica

The problems below include the exact dates and times Chesthetica created the compositions. Typically, human composers do not create more than one problem per day (in some cases only one per month). But with machines the time index is relevant: Chesthetica may compose ten or twelve on the same day.


White to play and mate in three
14/9/2014 3:16:47 PM
 

White to play and mate in three
17/9/2014 3:55:26 PM

White to play and mate in three
17/9/2014 4:47:29 PM
 

White to play and mate in three
20/8/2014 7:04:40 PM

White to play and mate in three
19/9/2014 8:14:57 AM
 

White to play and mate in three
24/9/2014 1:51:11 PM

White to play and mate in three
28/9/2014 3:40:07 AM
 

White to play and mate in three
11/10/2014 11:23:36 AM

White to play and mate in three
3 /9/2014 10:33:18 PM
 

White to play and mate in three
25/9/2014 6:21:59 AM

Instead of further details, since I am not a computer expert, let me study the products, since that is actually my field. The solutions to most of the problems are hidden and require a click to be revealed. This is in order to offer you the opportunity to try and solve them yourself, as an additionaly little benefit.

Here is the first Chesthetica problem we want to look at:

Chesthetica, 14/9/2014 3:16:47 PM

White to play and mate in three moves

The solution given is: 1.Qc3 Rh7+ 2.Bxh7. Let us tke a closer look. 1.Qc3 threatens both 2.Qg3 mate and 2.Nh3 mate. But there is no need at all for the e1 bishop, and its presence is in fact causing duals. It is a serious mistake. So we can simply remove it.

Furthermore in the starting position there are three possibilities for checking the white king, but none result in mates. That significantly helps the solving. On the top of that the key move takes away two flight squares of the black king. The remaining check opportunity only makes the problem one move longer, so it is virtually only a two moves problem. If we remove the bishop on e1 and move the one on g6 to h7, and we move the black rook from e7 to e8 we have the following position:

Now it is White to play and mate in two moves

The solution: 1.Qc3 Re3 (to counter both threats, 2.Qg3 and 2.Nh3 mate) 2.Qf6 mate. Much more satisfying, don't you think?

Here's a problem I created with similar material:

Pal Benko, 2015

White to play and mate in two moves

The solution 1.Ng1! [1.Ng5? Qb3!] 1..Bf3 2.Nh3 mate. There is a twin:

Pal Benko, 2015

White to play and mate in two moves

The solution 1.Ng5! [1.Nd2? f1N!] 1...Qb3 2.Qd4 mate.

Here is another work to show the provided checks:

Pal Benko, 2015

White to play and mate in two moves

The solution 1.Qb8! We can see two checks to the white king, but both KxRe3 and KxRd4 results discovered mate. After the key move it changes to NxRd4 and NxRe3 mate. We can find further ideas in this piece: the position is symmetrical, the axis is the b1-h7 diagonal. But the key is asymmetrical, that makes it interesting.

And let us see an item from the past.

Pal Benko, Chess Life 1980

White to play and mate in two moves

The solution 1.Qb4! A waiting move that gives two flight squares for the king to escape! Furthermore the possible moves of the black pawn give a theme to this miniature. Can a computer appreciate this? The answer is hardly ever possible.

Replay all the above problems

Select games from the dropdown menu above the board

About the author

Pál Benkö, 86, is a Hungarian-American chess grandmaster, openings theoretician, author and problemist. He became Hungarian champion when he was 20 and finished in first place (or tied for first place) in eight US Championships, a record: 1961, 1964 (in that year he also won the Canadian Open Chess Championship), 1965, 1966, 1967, 1969, 1974, 1975. Benko's highest achievements were playing in the Candidates Tournament with eight of the world's top players in 1959 and 1962. He qualified for the 1970 Interzonal tournament, the leaders of which advance to the Candidates. However, he gave up his spot in the Interzonal to Bobby Fischer, who went on to win the World Championship in 1972.

In addition to his success as a player, Benko is a noted authority on the chess endgame and a composer of endgame studies and chess problems. He is an over-the-board GM and also a FIDE IM of chess composition. The only other person we know who has these two titles is Jan Timman of the Netherlands.

– Part two of Pal Benko's article will follow soon –


Topics problems, Benko

World class chess grandmaster, author, and composer of endgame studies and problems. Benko qualified for the Candidates Tournament for the World Championship in 1959 and 1962, and for the 1970 Interzonal tournament, when he gave up his spot to Bobby Fischer, who went on to win the World Championship in 1972. Pal was born in 1928 and lives in Budapest, Hungary.
Feedback and mail to our news service Please use this account if you want to contribute to or comment on our news page service



Discuss

Rules for reader comments

 
 

Not registered yet? Register

jimliew jimliew 6/14/2016 06:20
Is there a point to this article?
kyi kyi 6/14/2016 01:14
Why did he give up his spot in the Interzonal to Bobby Fischer, who went on to win the World Championship in 1972 ? My impression is that Hungary was brutally suppressed under former communist rule when Hungarians revolted against former Soviet Union. The world chess was dominated by USSR chess after second world war. Pal Benko believed that Fischer was the only one and not him who could demolish Soviet chess hegemony. That might be the reason to revenge Soviet brutality of Hungarians, he gave up his spot to Bobby Fischer. Did Fischer thank and fully compensate him ? If it is not so, did he regret giving up his spot to him and missed the chance to become world chess champion ? I appreciate if Pal Benko can give me those answers.
MatPlus2 MatPlus2 6/14/2016 01:28
First of all, please try to find out what is considered beautiful in the art of chess problems. The criteria used by computers were valid some 300 years ago, and those "ideas" have already been composed by humans. You can choose some of the FIDE Albums (the best composition for a 3 year period) or recent Encyclopedia of Chess Problems_Themes and Terms by one of the world greatest experts Milan Velimirovic, and make sure that you really understand what a chess problem is in the first place!
keram1969 keram1969 6/14/2016 01:55
Great player and beautiful chess problems , especialy mat two bishops is interestig for me .
Qc3 tempo !!!
eltollo eltollo 6/14/2016 01:57
MatPlus2 suggests an interesting research track for Dr Alsan Iqbal: apply machine learning techniques to learn from the FIDE Albums what a beautiful study is, subsequently generate random positions and let the classifier decide if such a random position is a "beautiful study".
azlan azlan 6/14/2016 04:03
@eltollo: Unfortunately, AI funding bodies see little value in funding the development of a machine that basically just composes (even) high quality chess problems. There may be some interest in a new kind of AI technology *behind* such a program but it needs to have potentially (more important) applications in other areas. That usually means that, at first, it needs to be less specific or tailored to any particular task.
genem genem 6/14/2016 08:05
Nice analysis.

@kyi Benko sensed he had been caught in a rare fleeting moment where something very special could happen in he chess world, but only if he made a steep personal sacrifice. I strongly doubt Benko "regrets" his decision, because the heavenly stars then fell into alignment and the very special moment more than lived up to what Benko and others foresaw.
Bobby Fischer suffered from some degree of mental illness, and nothing short of extraordinary efforts by others could have helped Bobby overcome his daemons long enough to achieve the goal that we all wanted for Bobby.
Add the name of Jim Slater to those who helped make Spassky-Fischer 1972 happen.
Peter B Peter B 6/15/2016 01:53
A problem is that a computer does not have a good idea for what is "easy" or "hard" for a human to see. Some of the above problems are very simple (e.g. the K+R+P v K one), even though they have only one solution.
gmwdim gmwdim 6/15/2016 04:23
Benko was too reserved in his criticism of the machine problem in question.
azlan azlan 6/15/2016 08:03
@Peter B: There are probably ways to program a computer to filter for "difficulty" in terms of the search depth and unlikeliness of the key move in the main line of the solution. The other criteria described in the article are likely programmable as well. In fact, there could be highly specialized composers of three-movers, four-movers, studies etc. just like human composers tend to specialize. However, these are whole other projects in themselves.

Again, the AI community and funding bodies are typically not interested in funding such things (I have tried) because they are pretty much useless for any other purpose. A highly specialized composer of a certain class of endgame study, for example, would probably be useless (by master composer standards) even in composing three-movers, much less applicable in other more important areas that could benefit from AI. Just as Deep Blue, even in its day and for all the money, time and effort that went into it, was pretty much useless for any other purpose aside from playing chess well enough to beat the world champion (which was perhaps the main objective of IBM anyway).
satman satman 6/15/2016 10:54
@azlan
It was Oscar Wilde who opined 'All art is quite useless'.
By this criteria Problem Chess is the purest art form - even more so because there's little or no money involved.
The great masterpieces of chess composition are at the same time both priceless and worthless.
Will computers ever be able to match these achievements?
Obviously the answer must be Yes - but as you state, for the time being it's the very uselessness of the form which is putting off the dreaded day.
1