An oracle can be used to compress strings and to test sequences for randomness. A set is low for a relativizable notion if, used as an oracle, it does not change the notion. For prefix-free complexity and Martin-Löf randomness, the low sets are exactly the K-trivial sets. This lesson proves the equivalence of three lowness notions, by the hungry sets argument of Hirschfeldt, Nies and Stephan, and states the theorem of Nies and Hirschfeldt that connects them to K-triviality.
3.1 Relativized complexity and randomness
An oracle prefix-free machine is an oracle machine such that, for every oracle , the domain of is prefix-free. The constructions of Prefix-free complexity relativize uniformly: there is an oracle prefix-free machine such that, for every oracle prefix-free machine , there is with for every oracle . We write , and for the class of sequences passing every Martin-Löf test relative to , the tests whose components are uniformly . The results of the course on Martin-Löf randomness relativize, by the same proofs; in particular if and only if , and .
If , then , since a machine with oracle can simulate ; in particular . So oracles can only make complexity smaller and randomness harder.
For a string , we write when the computation of on input with oracle halts with output and its use is exactly , in the sense of Definition 1.6 of “Computation relative to an oracle”. The set of triples with is c.e., and for every oracle and every , there is exactly one with , namely for the use of the computation.
3.2 Three lowness notions
Let be a set.
- is low for K if for every string .
- is low for Martin-Löf randomness if .
- is a base for randomness if for some .
Every computable set has the three properties. For a non-computable set , the sequences computing form a null class (Sacks); a base for randomness is a set computed by some sequence that is nevertheless random relative to .
- If is low for K, then is K-trivial.
- If is low for K, then is low for Martin-Löf randomness.
- If is low for Martin-Löf randomness, then is a base for randomness.
- The sets that are low for K are closed downward under .
- With oracle , the string is computed from , so . If is low for K, then .
- for every , since a test without oracle is a test relative to . Conversely, let , so that . If is low for K, then , and by the relativized Levin–Schnorr theorem.
- By the Kučera–Gács theorem (Kučera, Gács), some computes , and .
- If and is low for K, then .
3.3 The hungry sets theorem
Every base for randomness is low for K.
Let and a Turing functional with . For a string and a string , write if for every .
The construction. Fix a parameter . For each triple with we enumerate a finite or infinite set of strings , the hungry set of the triple, which aims at a measure . Each triple is first considered at the stage at which the computation is found. At each stage , the triples considered so far are processed in turn. For the triple , each string of length such that
- ,
- is disjoint from the classes of all hungry sets , in their current state, and
- ,
is added to , the strings being examined in lexicographic order. At the end of the processing of the triple, if and no request has yet been made for this triple, the request is enumerated into a set .
At stage , every string in a hungry set has length at most , so for the cylinder is either contained in or disjoint from each class ; condition 2 is decidable. By condition 2, the classes of distinct hungry sets are pairwise disjoint, and by condition 3, .
is a bounded request set. A request is made for a triple only when its hungry set has measure at least , the weight of the request. Since the hungry sets are disjoint, the total weight of is at most .
A test relative to . Let be the union of the classes over the triples with , for the construction with parameter . It is , uniformly in . For each there is exactly one triple with , so
So is a Martin-Löf test relative to , and since , there is with . Fix this .
Every hungry set of a true computation becomes half full. Let with , and suppose that the hungry set never reaches measure . Since , there is a stage , after the triple is first considered, with and . Let , and consider the processing of the triple at stage . Condition 1 holds for . If were contained in a hungry set , then some was added to , so and, since , ; then , a contradiction. So condition 2 holds. Condition 3 holds since . So is added to , and , again a contradiction.
Conclusion. For every with , the triple with has its hungry set half full, and the request belongs to . By the KC theorem applied to the c.e. bounded request set , . Taking a shortest description of relative to gives .
The hungry sets convert descriptions relative to into descriptions without oracle. A description of with oracle prefix is paid for by measure: the hungry set eats the oracles that compute through , until it has eaten . If it starves, is in a hungry set of a true computation, which is impossible when is chosen so that avoids the -th component of a test relative to . The parameter is not known effectively, but this is harmless: for each , is a c.e. bounded request set, and only the existence of a good is needed.
For every set , the following are equivalent: is low for K; is low for Martin-Löf randomness; is a base for randomness.
By Proposition 3.2 (2) and (3), and Theorem 3.3.
In particular, by Proposition 3.2 (1), all these sets are K-trivial, hence .
3.4 K-triviality and lowness
The converse of Proposition 3.2 (1) is the central theorem on K-triviality.
Every K-trivial set is low for K.
The proof, known as the golden run construction, is a priority argument of considerable length, and is admitted here; see Nies (2005), or Nies, Computability and Randomness, Section 5.4, and Downey and Hirschfeldt, Algorithmic Randomness and Complexity, Chapter 11. Another proof, which avoids the golden run, goes through the theorem of Bienvenu, Day, Greenberg, Kučera, Miller, Nies and Turetsky (2014) that every c.e. K-trivial set is computable from a Martin-Löf random sequence that does not compute ; it also uses Theorem 3.3 and the c.e. trees of K-trivial sets. Granting the theorem, the following consequences are immediate.
For a set , the following are equivalent:
- is K-trivial;
- is low for K;
- is low for Martin-Löf randomness;
- is a base for randomness.
The K-trivial sets form an ideal of the Turing degrees: they are closed downward under , and under .
(1 ⇔ 2) by Theorem 3.5 and Proposition 3.2 (1); the other equivalences by Corollary 3.4. Closure downward holds for low for K sets by Proposition 3.2 (4), and closure under join for K-trivial sets by Proposition 1.5.
Downward closure is not apparent from the definition of K-triviality, and closure under join is not apparent from the definition of lowness for K; the equivalence yields both.
3.5 Exercises
Show that every set that is low for K is low, that is, . Deduce, using Theorem 3.5, that no K-trivial set is Turing complete.
Hint
For , the use of the computation gives a prefix of with a short description relative to .
Solution
Let be low for K; it is K-trivial, hence . Let be the partial -computable function with when with use . Then for , since is computable in . By Proposition 3.4 of “Prefix-free complexity” (2), . Since is low for K, there is with for every , where is computable.
With oracle , decide as follows: for each string of length at most , decide whether converges; if it does, with output , decide whether , which is possible since , and whether , a question. Answer yes if some passes all three tests. If , then a shortest description of passes them. Conversely, if and , then by the use principle. So .
A complete set has , which is not computable in ; so a K-trivial set, which is low for K by Theorem 3.5, is not complete.
Show without Theorem 3.5 that a base for randomness is not Turing complete.
Solution
By Theorem 3.3, is low for K, and by the first exercise it is low, hence not complete.
3.6 Notes and references
Lowness for Martin-Löf randomness was studied by Zambella (1990) and by Kučera and Terwijn (1999); the latter built a non-computable c.e. set low for randomness. Lowness for K is due to Muchnik, who built a non-computable c.e. set low for K. Nies (2005) proved that low for K and low for randomness coincide, and, with Hirschfeldt, that both coincide with K-triviality. The notion of base for randomness and the hungry sets argument are due respectively to Kučera (1993) and to Hirschfeldt, Nies and Stephan (2007).
- A. Nies, Lowness properties and randomness, Advances in Mathematics 197 (2005), 274–305.
- D. R. Hirschfeldt, A. Nies and F. Stephan, Using random sets as oracles, Journal of the London Mathematical Society 75 (2007), 610–622.
- A. Kučera, On relative randomness, Annals of Pure and Applied Logic 63 (1993), 61–67.
- A. Kučera and S. A. Terwijn, Lowness for the class of random sets, Journal of Symbolic Logic 64 (1999), 1396–1402.
- L. Bienvenu, A. R. Day, N. Greenberg, A. Kučera, J. S. Miller, A. Nies and D. Turetsky, Computing K-trivial sets by incomplete random sets, Bulletin of Symbolic Logic 20 (2014), 80–90.