ECM2418 Computer Languages and Representation

Question 1

Haskell inbuilt functions to use

Answer 1

filter
foldr
map
even
odd
(++)
sel
not

Question 2

polymorphic type of conditional functor (even, odd) haskell

Answer 2

(a -> Bool)

Question 3

polymorphism syntax haskell

Answer 3

function :: a..

Question 4

list split haskell

Answer 4

(x:xs)

Question 5

list split Prolog

Answer 5

([A|AS])

Question 6

Prolog predicates to remember

Answer 6

reverse
between
findall
permutations

Question 7

(EXAMPLE) what’s stopping this predicate from working

alpha(XS).

Answer 7

it’s missing base cases:
alpha([]).
alpha([_]).

Question 8

Functional programming

Answer 8

• Programs are represented by expressions
• Computation is implemented by reduction
• The foundation is the lambda calculus
• Functions akin to mathematics functions

Question 9

Logic Programming

Answer 9

• Programs are represented by clauses
• Computation is represented by proof
• The foundation is the first-order logic

Question 10

Finite State Machines

Answer 10

Simple idea occurring throughout computer science
- Programs are represented by state machines
- Computation is represented by transitions

Question 11

tuples in haskell

Answer 11

(a1,a2,a3)
can have different types in one tuple

Question 12

Recursive definitions in haskell

Answer 12

Contains a base case that can act as a halt for the function to return a value

Question 13

Guards in Haskell

Answer 13

|

a boolean expression that must be true for the equation to apply

acts similarly to if and else

Question 14

guard keywords

Answer 14

where
otherwise

Question 15

generate-and-select design pattern

Answer 15

the generator constructs a large number of values that might be solutions to a problem

the selector filters out those values that are solutions to the problem

Question 16

explanations of the productivity paradox

Answer 16

1) Uneven/concentrated gains
2) Implementation lags
3) Mismeasurements

Question 17

uneven/concentrated gains

Answer 17

One explanation of the productivity paradox is an uneven/concentrated distribution of gains - the gains are in few productive firms and sectors with limited weight int the overall economy

Question 18

Implementation lags

Answer 18

Another explanation of the productivity paradox is implementation lags - it takes considerable time for new technologies to achieve critical mass, or for necessary complementary ones appear. There are gains, but it takes a long time time

Question 19

Mismeasurement

Answer 19

adopting new technologies can lead workers to move from more productive adopting sectors to less productive ones, and so to negligible aggregate productivity growth

Question 20

Behind the scenes, Artificial Neural Networks (ANNs) work as follows:

Answer 20

1) The core ANN is trained on text from the Internet to respond to a prompt with a list of most probable next words after the prompt

2) The core ANN is tweaked by scoring its responses to sample queries. A second ANN is trained on these scores to predict the most likely one to be assigned to a prompt

3) The second ANN is used in reinforcement learning to adjust the weights in the core ANN so that its outputs are even more likely to satisfy humans

4) Sometimes, data from user responses to LLM responses is fed back to fine-tune for still better results

Question 21

Among the concerns when using Large Language Models are:

Answer 21

• copyright
• education
• code quality
• code security
• not an expert

Question 22

Copyright and LLMs

Answer 22

although the code produced is the melding of that form many sources, there may still be some licensing problems

Question 23

Education and LLMs

Answer 23

good solutions to many assignments can be completed directly, reducing their challenge and worth

Question 24

code quality and LLMs

Answer 24

There is no guarantee of code quality when using a large language model, and a code review will need to be conducted before code is deployed

Question 25

Code security and LLMs

Answer 25

There is no guarantee of code security when using a LLM, and a security audit will need to be conducted before code is deployed

Question 26

LLMs are not experts

Answer 26

it does not know that it does not know, able to confabulate is necessary, mixing high-quality text with low-quality rubbish

Question 27

List comprehensions

Answer 27

[ | ] used to test and transform a list

Question 28

conditionals in haskell

Answer 28

case _ of _ -> or if _ then _ else

Question 29

cody copying is

Answer 29

error-prone inefficient

Question 30

Backtracking in Prolog

Answer 30

If any goal fails, the Prolog system tries to find another way of satisfying the most recently satisfied previous goal Alternative proofs are requested with NEXT, forcing backtracking Successive permutations of a clause can be achieved via backtracking

Question 31

atoms in Prolog

Answer 31

An atom is a constant that does not have a numerical value. It begins with a lowercase letter, or is quoted It is possible to split atoms into lists of ASCII codes, and to make atoms from lists of ASCII codes This makes Prolog and other logical languages a great fit for Natural Language Processing

Question 32

Prolog's Closed World Assumption (CWA)

Answer 32

Anything that cannot be proven to be true based on the facts and rules in the program is assumed to be false For example, if a fact likes(mary, pizza) exists but likes(mary, sushi) does not, Prolog assumes likes(mary, sushi) is false.

Question 33

Declarative Semantics in Prolog

Answer 33

Logic programs describe what is true, not necessarily how to compute it. Example: Writing ascending([X, Y | XYS]) :- X =< Y, ascending([Y | XYS]). specifies the property of an ascending list, not how to iterate through the list procedurally.

Question 34

Horn Clauses Prolog

Answer 34

Programs are expressed as a set of Horn clauses, which are a type of logical implication: Facts: Unconditionally true statements, e.g., parent(john, mary). Rules: Conditional statements, e.g., ancestor(X, Y) :- parent(X, Z), ancestor(Z, Y).

Question 35

Resolution Prolog

Answer 35

The main inference mechanism, where Prolog tries to resolve goals by matching them against facts or rules.

Question 36

Unification Prolog

Answer 36

Prolog matches variables with constants or other variables to make clauses work. It ensures terms are syntactically compatible.

Question 37

Why might one discourage the use of the assert and retract predicates in prolog

Answer 37

Assert adds a predicate to the running program, and the retract predicate removes. They make reasoning about the program harder The program being run changes over time Perhaps less efficient code gets created and destroyed on program run-time

Question 38

Unlike Haskell programs, Prolog programs are not strongly typed. What does this mean, and why might it cause problems?

Answer 38

This means that due to programmer misunderstandings or errors, predicates may be applied to arguments of unexpected types, predicate calls may fail due to unification errors, incorrect results. This failure will be silent and hard to debug

Question 39

X // Y vs X / Y

Answer 39

X // Y represents the integer quotient of X and Y, while X / Y can compute a floating point result

Question 40

arithmetic functions abs(N, X) sin(N, X) cos(N, X) log(N, X) sqrt(N, X)

Answer 40

absolute value of N sine of N in radians cosine of N in radians natural logarithm of N square root of N

Question 41

Relational operators X =:= Y X =\= Y X > Y X >= Y X < Y X =< Y

Answer 41

X and Y are same value X and Y are not same value X is greater than Y X is greater than or equal to Y X is less than Y X is less than or equal to Y

Question 42

Unification operators X = Y X \= Y

Answer 42

X and Y can be unified X and Y cannot be unified

Question 43

Which of these terms unifies: * 42 and [42]; * anna and X; * woman(anna) and woman(X); * [’A’,’B’,’C’] and [P, Q, R]; * loves(X, jo) and loves(jim, X)

Answer 43

* anna and X; * woman(anna) and woman(X) * [’A’, ’B’, ’C’] and [P, Q, R] * loves(X, jo) and loves(jim, X)

Question 44

logical negation operator

Answer 44

\+ X means Not X Note the closed-word assumption — what is stated is true; what is not stated is false. Also note negation as failure — anything that cannot be proved true is false

Question 45

The Cut Predicate, '!', in Prolog

Answer 45

prevents backtracking Although backtracking is a fundamental part of the process by which goals are satisfied, it can sometimes be too powerful.

Question 46

Classical Expert System

Answer 46

a database of elementary facts about the domain; some domain rules for knowledge acquisition; some inference rules for reasoning from data; a facility for generating explanations;

Question 47

examples of classical expert system

Answer 47

Dendral MYCIN PROSPECTOR SHRDLU

Question 48

Dendral, 1965

Answer 48

generates all possible arrangements of a set of atoms, and searches through it for likely ones, consistent with the rules of chemical valence. The heuristics employed are based on judgment and chemical knowledge — experience and intuition.

Question 49

MYCIN, 1972

Answer 49

designed to assist physicians in the diagnosis of bacterial infections. Knowledge is encoded as 350 decision rules which embody the clinical decision criteria of infectious disease experts

Question 50

PROSPECTOR, 1976

Answer 50

In expert systems, a blackboard architecture is one where a shared “blackboard” is iteratively updated by a diverse group of knowledge sources, starting with a problem specification and ending with a solution. The Prolog database readily supports this architecture.

Question 51

Static Predicates

Answer 51

cannot be modified as the program runs — by default, predicates are static ones.

Question 52

Dynamic Predicates

Answer 52

can be modified as the program runs — predicates may be declared as dynamic ones. dynamic(droid/1).

Question 53

Adding clauses to the database

Answer 53

asserta(X) - add a clause X at the beginning of the database; assertz(X) - add a clause X at the end of the database; assert(X) - add a clause X at the beginning of the database.

Question 54

Deleting clauses from the database

Answer 54

retract(X) — remove clause X from the database; retractall(X) — remove all facts/clauses that unify with X from the database

Question 55

The Fifth Generation Project

Answer 55

The Fifth generation Project was intended to: - make Japan a leader in the computer industry by developing knowledge information processing systems built using logic programming - stimulate original research, making the results widely available, so refuting accusations that Japan exploits knowledge from abroad without contributing any of its own,

Question 56

Logic programming was chosen for the Fifth Generation Project because it unifies innovations in:

Answer 56

1 software engineering; 2 databases; 3 artificial intelligence; 4 computer architecture.

Question 57

Software Engineering and Logic programming

Answer 57

In the field of software engineering, logic programs can serve as executable specifications, that can be made more efficient by program transformation

Question 58

databases and logic programming

Answer 58

In the field of databases, logic programs have been shown to be adequate both as a query language, and for describing integrity constraints

Question 59

Artificial Intelligence and Logic programming

Answer 59

In the field of artificial intelligence, logic programs naturally implement rules of the form A if B1 and . . . and Bn

Question 60

Computer Architecture and Logic Programming

Answer 60

Logic programs provide a way to overcome the Von Neumann Bottleneck through single assignment

Question 61

Inference Machines

Answer 61

Several hundred Personal Sequential Inference Machine (PSI) workstations were built and installed. The largest Parallel Inference Machine (PIM), PIM/p had 64 × 8 processing elements

Question 62

Three "classical forms of parallelism in Prolog"

Answer 62

unification-parallelism or-parallelism and-parallelism

Question 63

Unification Parallelism

Answer 63

Unification parallelism arises during the unification of the arguments of a goal with the arguments of a clause head with the same name and arity. Unification parallelism unifies the arguments of complex structures in parallel.

Question 64

Or-parallelism

Answer 64

Or-Parallelism arises when a goal may be unified with more than one clause in parallel. Or-Parallelism explores the search space generated by the presence of multiple clauses in parallel

Question 65

regular expression algebra uses 3 different operations

Answer 65

closure concatenation union

Question 66

Closure, L∗

Answer 66

represents the set of those strings that can be formed by taking any number of strings from L, and concatenating them has the highest precedence, applying to the smallest sequence of symbols to its left

Question 67

Concatenation, LM

Answer 67

the set of strings that can be formed by taking any string in L and appending any string in M second highest priority - expressions that are juxtaposed are grouped together

Question 68

Union, L + M

Answer 68

set of strings either in L or M or both lowest priority - expressions may be grouped by unions in any order.

Question 69

0* + 1

Answer 69

any number of 0s or a 1

Question 70

101*

Answer 70

10 and any number of 1s

Question 71

10 + 01

Answer 71

10 or 01

Question 72

0(10)*

Answer 72

0 and any number of 10s

Question 73

Five advantages that the Japanese Fifth Generation Project saw in using Prolog programs as specifications were:

Answer 73

(i) specifications are naturally described by logical relation- ships [1 mark]; (ii) logical relationships are naturally described by Prolog pro- grams [1 mark]; (iii) specifications and Prolog programs are therefore always consis- tent [1 mark]; (iv) Prolog specifications/programs are executable, albeit rather ineffi- ciently [1 mark]. (v) Prolog specifications/programs can be transformed into more efficient ones with machine assistance [1 mark].