Almost-Smooth Histograms Overview

• The paper presents an almost-smooth histogram framework that achieves a (2+O(ε))-approximation for all bounded, non-decreasing subadditive functions.
• It uses a bucket-based maintenance procedure to provide logarithmic space and update time for sliding-window approximations in streaming algorithms.
• The framework broadens applicability to symmetric norms, graph parameters, and discrete density estimations, offering near-universal data reduction tools.

An almost-smooth histogram is a data structure and analytical device that extends the classical smooth-histogram framework to a substantially broader class of functions, particularly those which are subadditive but fail to satisfy the strict requirements of smoothness. This paradigm plays a central role in streaming algorithmics, allowing efficient, space-bounded, approximate computation over sliding windows in data streams for a wide variety of objectives—including those for which previous smooth-histogram approaches cannot apply. The almost-smooth histogram framework, as developed by Krauthgamer and Reitblat, yields $(2+O(\epsilon))$-approximation guarantees for all bounded, non-decreasing subadditive functions and provides a near-universal data reduction and approximation toolkit in the streaming and graph analytics domain (Krauthgamer et al., 2019). Related but distinct are methodologies in density estimation for discrete distributions with “almost-smooth” histograms—kernel smoothers with positive mass everywhere but preserving the discrete structure of the data (e.g., the CMP kernel) (Huang et al., 2020), and regularized spline-based fits for binned continuous data (Goulko et al., 2017), which may—by analogy—be informally regarded as “almost-smooth” reconstructions.

1. Formal Framework: Smoothness and Almost-Smoothness

The classical Braverman–Ostrovsky smooth-histogram framework requires a function $f$ (on streams) to be $(\alpha,\beta)$-smooth, meaning that a suffix which nearly retains the full value of a function cannot, when prefixed or suffixed, suddenly lose this property. Specifically, when $f$ evaluated on segment $B$ is close to $f$ on $AB$, this stability must be maintained upon further extension. For a non-negative stream function $f$, smoothness is:

• $\beta(\delta)\leq \delta$ for all $\delta\in(0,1)$;
• Whenever $f$0 for disjoint $f$1, then for every $f$2,

$f$3

However, many natural functions—particularly subadditive, bounded, non-decreasing set functions such as matching size and symmetric norms—do not satisfy this strict form of stability. To accommodate these, the framework introduces $f$4-almost-smoothness:

A function $f$5 is $f$6-almost-smooth if it is

• $f$7-left-monotone: $f$8 for any disjoint $f$9;
• $(\alpha,\beta)$0-suffix-stable: $(\alpha,\beta)$1 for any disjoint $(\alpha,\beta)$2 with $(\alpha,\beta)$3.

Every bounded, non-decreasing, subadditive function is $(\alpha,\beta)$4-almost-smooth. This includes maximum matching size, symmetric norms, vertex cover size, and combinatorial graph parameters not covered by the original smooth-histogram machinery (Krauthgamer et al., 2019).

2. Construction and Maintenance of Almost-Smooth Histograms

The almost-smooth histogram data structure generalizes the bucket-based maintenance procedure of smooth histograms. The system keeps an ordered sequence of $(\alpha,\beta)$5 buckets for window size $(\alpha,\beta)$6 and approximation parameter $(\alpha,\beta)$7. Each bucket $(\alpha,\beta)$8 stores the state of an insertion-only algorithm $(\alpha,\beta)$9 for $f$0. Upon arrival of a new item, each bucket’s algorithm is updated, a new bucket is created, and buckets are merged or discarded if adjacent values are close (factor $f$1).

Update Procedure:

1. On item $f$2 arrival, create $f$3, instantiate $f$4.
2. For all $f$5, update $f$6 with $f$7 (so $f$8).
3. Scan for the largest $f$9 with $B$0. Delete buckets $B$1 if found.
4. If $B$2 is older than $B$3, discard $B$4.

Query Procedure:

• If $B$5 (current window), output $B$6.
• Otherwise, output

$B$7

for $B$8-almost-smooth $B$9.

This bucket management ensures a logarithmic-in-window, inverse-linear-in-epsilon complexity blow-up over the underlying insertion-only streaming algorithm (Krauthgamer et al., 2019).

3. Approximation Guarantees and Invariants

For $f$0-almost-smooth $f$1 with a $f$2-approximate insertion-only streaming algorithm $f$3, the sliding-window almost-smooth histogram produces a

$f$4

with $f$5 overhead in space and update time (Krauthgamer et al., 2019).

Key invariants:

• The buckets form a suffix chain: $f$6.
• Adjacent buckets differ by at least an $f$7 fraction in their $f$8-estimates.
• Monotonicity and almost-smoothness ensure the true function values in the window are bracketed by the bucket estimates up to the $f$9 constants.

This methodology requires no explicit storage of streamed items, only algorithm states and counters, and directly generalizes and subsumes the classical smooth-histogram for $AB$0.

4. Applications: Streaming Algorithms Beyond Smoothness

The almost-smooth histogram framework directly enables sliding-window (i.e., fixed-length suffix) approximations for numerous objectives previously inaccessible to streaming algorithms:

• Symmetric norms on frequency vectors: Every symmetric norm $AB$1 (e.g., top-$AB$2 norms) is subadditive and monotone; a $AB$3-approximation in $AB$4 space is obtained, where $AB$5 is the modulus of median concentration.
• Schatten 4-norm: For streamed matrices, the Schatten 4-norm is $AB$6-almost-smooth; the approach yields a $AB$7-approximate sliding-window algorithm in $AB$8 space.
• Graph combinatorial objectives: Maximum submodular matching, minimum vertex cover, and maximum $AB$9-cover size—all subadditive—admit sliding-window constant-factor approximations using the insertion-only algorithms as black boxes (Krauthgamer et al., 2019).

Artificial families of functions parametrized by almost-smoothness constant $f$0 are also constructed, demonstrating the strictness and granularity of the framework’s inclusivity.

5. Contrast with Related “Almost-Smooth” Histogram paradigms in Estimation

Separately, the term “almost-smooth histogram” is used in discrete distribution estimation, where kernel smoothing is performed over count histograms to generate estimators with no zero-probability gaps and strictly positive support, but which do not admit a closed-form smooth density (Huang et al., 2020). The mean-parametrized Conway-Maxwell-Poisson (CMP) kernel defines a family of discrete, infinitely-supported, second-order kernels with variance $f$1 and mean $f$2; the smoothed estimator

$f$3

assigns positive mass at all counts, infilling gaps within and beyond the observed range. Automated bandwidth selection is handled either by Kullback–Leibler divergence minimization against reference Poisson or Negative-Binomial fits ($f$4), or by leave-one-out cross-validated predictive likelihood ($f$5). The resulting estimator achieves mean integrated squared error rates $f$6 and strictly dominates ordinary histograms in both simulation and real data (Huang et al., 2020).

6. Extensions, Limitations, and Impact

Almost-smoothness generalizes the strict smooth-histogram paradigm by tolerating a bounded deterioration factor $f$7 in suffix-stability, thereby encompassing all bounded subadditive objectives and some monotone submodular graph parameters. The framework’s main implication is that any $f$8-approximate insertion-only streaming algorithm for such an objective induces—essentially “for free”—a sliding-window $f$9-approximate algorithm with logarithmic overhead in window size. This unlocks a large class of efficient data stream computations previously thought inaccessible outside of smooth function classes (Krauthgamer et al., 2019).

A notable boundary is that functions with arbitrary (unbounded) non-smoothness are not included; subadditivity and bounded monotonicity are required. Empirical kernel-based “almost-smooth” reconstructions for discrete distributions share the spirit of smoothing but are analytically and algorithmically distinct (Huang et al., 2020). Analogous spline-based bin hierarchy methods in numerical density estimation (Goulko et al., 2017) further illustrate the broad relevance of “almost-smooth” concepts to the construction of robust, artifact-free reconstructions from binned or sampled data.

7. Comparison Table of Key Approaches

Framework	Function Classes Covered	Sliding-Window Approximation	Structural Guarantee
Smooth histogram [Braverman–Ostrovsky]	$\beta(\delta)\leq \delta$0-smooth (e.g., $\beta(\delta)\leq \delta$1 norms, some polynomials)	$\beta(\delta)\leq \delta$2	Strict suffix-stability ($\beta(\delta)\leq \delta$3)
Almost-smooth histogram (Krauthgamer et al., 2019)	$\beta(\delta)\leq \delta$4-almost-smooth ($\beta(\delta)\leq \delta$5-left-monotone, $\beta(\delta)\leq \delta$6-suffix-stable)—includes all bounded subadditive, monotone objectives; combinatorial graph functions	$\beta(\delta)\leq \delta$7	Bounded suffix-stability ($\beta(\delta)\leq \delta$8)
CMP kernel smoothing (Huang et al., 2020)	Discrete pmf, all counts with positive mass; no streaming focus	Consistent estimator; nonzero gap-filling	Discrete, positive, 2nd order kernel

Almost-smooth histograms thus occupy a crucial position in modern streaming and statistical algorithmics, permitting efficiently maintainable, space-bounded, accurate approximation for broad classes of non-smooth or subadditive functions that naturally arise in data analysis and graph computation.