Deciding Monotone Duality and Identifying Frequent Itemsets in Quadratic Logspace

The monotone duality problem is defined as follows: Given two monotone formulas f and g in iredundant DNF, decide whether f and g are dual. This problem is the same as duality testing for hypergraphs, that is, checking whether a hypergraph H consists of precisely all minimal transversals of a simple hypergraph G. By exploiting a recent problem-decomposition method by Boros and Makino (ICALP 2009), we show that duality testing for hypergraphs, and thus for monotone DNFs, is feasible in DSPACE[log^2 n], i.e., in quadratic logspace. As the monotone duality problem is equivalent to a number of problems in the areas of databases, data mining, and knowledge discovery, the results presented here yield new complexity results for those problems, too. For example, it follows from our results that whenever for a Boolean-valued relation (whose attributes represent items), a number of maximal frequent itemsets and a number of minimal infrequent itemsets are known, then it can be decided in quadratic logspace whether there exist additional frequent or infrequent itemsets.

💡 Research Summary

The paper tackles the classic monotone duality problem, which asks whether two monotone Boolean formulas f and g (given in irredundant DNF) are dual to each other. Duality means that the set of satisfying assignments of f is exactly the complement of the set of satisfying assignments of g. This combinatorial question is equivalent to testing whether a hypergraph H consists precisely of all minimal transversals (also called hitting sets) of another hypergraph G. The problem appears in many areas, notably in database theory (query containment), knowledge discovery, and frequent‑itemset mining.

Historically, monotone duality has been placed in co‑NP and even shown to be co‑NP‑complete under certain restrictions, while the best known upper bound for the general case was PSPACE. Consequently, practical algorithms either required exponential space or relied on heuristics without formal guarantees.

The breakthrough of this work lies in applying the problem‑decomposition framework introduced by Boros and Makino (ICALP 2009) to the duality setting. Boros‑Makino’s method recursively splits a hypergraph into smaller sub‑instances while keeping the amount of information stored at each recursion level bounded by a logarithmic number of bits. The key properties of the decomposition are: (1) the recursion depth is O(log n), where n is the size of the input (number of variables plus the total length of the two DNFs); (2) each recursive call needs only O(log n) bits to describe the current candidate transversal; and (3) the branching factor is limited, so the total space consumption never exceeds O(log² n) bits.

The authors translate the monotone duality test into a hypergraph‑transversal test. They first convert the two DNFs into hypergraphs G and H by treating each clause as a hyperedge. Then, using a compact bit‑vector representation for variable sets, they run the Boros‑Makino recursion: a candidate transversal is selected, its presence in H is checked in logarithmic space, and if it fails, the candidate is split into two smaller candidates according to the decomposition rules. The recursion proceeds until either a counterexample is found (showing non‑duality) or all candidates are verified (establishing duality).

Complexity analysis shows that each recursion level performs only polynomial‑time operations (set inclusion tests, bit‑vector manipulations), while the depth contributes a factor of O(log n). Because the algorithm stores only the current candidate and a constant amount of bookkeeping information, the total space is bounded by DSPACE