Reference-distribution ambiguity: is not defined in the paper. If it is an arbitrary probability mass function, the printed inequality is false. For instance, choose and . Every empirical distribution equals , making the printed left side one bit and its right side zero. This is a counterexample for every .
Under the consistent interpretation , the preceding leave-one-out identity for empirical distributions and convexity of relative entropy in its first argument give, for every sample,
Deleting any one observation from independent and identically distributed random variables leaves independent observations with probability mass function . Consequently each has the same distribution as . They need not be independent of one another. Taking expected values and using linearity of expectation therefore proves
Zero coordinates of are never sampled, almost surely, so the calculation can be restricted to its probability support. In particular the expected values are finite on this finite alphabet.
The discrepancy cannot be removed by silently choosing any fixed reference. Whenever is positive on the probability support of , direct subtraction and give the expected relative entropy decomposition
Thus the valid arbitrary-reference version has on both sides; the printed mixed-reference version has an additional term whose sign prevents the proposed conclusion.
Index correction: the printed definition only permits , although the required average includes . Extend the same deletion formula to . Without that extension the final summand is undefined; there is no new sample or different estimator to supply.
For a fixed outcome of the sample and a fixed symbol , put . The leave-one-out empirical distribution then satisfies
Dividing by gives . Each of the indicator functions was counted exactly times before division by . This leave-one-out identity for empirical distributions is deterministic; independence plays no part in this part.
For probability mass functions on the finite alphabet, take logarithms to base two and define the relative entropy by
A term with is zero, including ; a term with is . Fix and write . For ,
Thus is a convex function on by continuous extension to zero. For , applying this scalar convexity to each coordinate and summing proves
If has zero coordinates, the finite-valued domain consists of probability mass functions supported on ; it is convex and the same proof applies there. Outside that domain the right side is infinite whenever an offending probability mass function has a positive mixture weight, so the extended-valued convexity assertion still holds. At mixture weights zero or one the inequality is an identity, without needing the expression .
Here the asserted bound is actually an identity, so equality always holds whenever the displayed quantities are defined. Expand the same mutual information in two orders:
Equating these expressions and using symmetry of mutual information and conditional mutual information gives
There is no further independence condition. Treating this part as a strict inequality would obscure its cancellation.
The chain rule for information entropy converts the two differences into conditional entropies. Their gap, in the order right minus left, is
This is additional conditioning reduces entropy. The relative entropy representation of conditional mutual information identifies exactly when that gap vanishes: equality holds if and only if and have conditional independence given . Knowledge of then changes no conditional probability of after is known.
Expand the two observations in order using the chain rule for mutual information:
The remaining conditional mutual information is
Relative entropy nonnegativity proves the inequality. Equality requires every positive-weight relative entropy to vanish, which is equivalent to for every of positive probability. Thus equality holds exactly when and have conditional independence given . No unconditional independence is required.
The chain rule for conditional entropy gives
Indeed, for every conditioning value of positive probability, each summand is nonnegative. Its sum is zero exactly when the conditional probability mass function is concentrated at one value. Averaging over therefore proves both the inequality and its equality criterion: equality holds exactly when almost surely for some function . Conditioning values of probability zero impose no restriction. This is a statement about discrete random variables; a differential entropy need not be nonnegative.
For the fundamental representation and the two-index symmetric representation, the Dynkin indices are and . Therefore the specified discrete gauge anomaly condition for is
All ordered solutions are
Only and give a faithful action on this pair of Weyl fermions; faithfulness was not imposed. As in solution, cancellation here concerns the stipulated discrete gauge anomaly, not cancellation of the pure cubic SU(3) gauge anomaly. The integer normalization matters: one must not replace by a half-integer trace index inside an unchanged modular congruence.
Use the integer normalization from solution. For two fundamental representations, the specified discrete gauge anomaly coefficient is . Consequently all solutions in the finite cyclic group of charges are
There are ordered charge pairs. The trivial pair is included: the problem does not require a faithful action. If a faithful action on these Weyl fermions were additionally required, one would retain precisely the coprime to .
This cancels the discrete gauge anomaly defined in the paper. Two left-handed fundamental representations by themselves have a nonzero pure cubic SU(3) gauge anomaly; a complete gauge theory would need additional fermions to cancel it. That separate condition does not alter the requested discrete congruence.
The two-index symmetric representation of SU(3) is , with Dynkin index . Thus copies have , and the supplied beta function is
The strict sign condition for asymptotic freedom is , giving
Indeed , while .
The Dynkin index is additive over a direct sum of group representations, so copies of the fundamental representation have . With , the supplied beta function becomes
For positive gauge coupling , asymptotic freedom at this order requires , hence . For nonnegative integer ,
At the one-loop coefficient vanishes, so this formula alone does not establish asymptotic freedom. This count follows the scalar multiplicity convention of the printed beta function; changing how a real or complex scalar field is counted requires changing its coefficient too.
Now the doublet invariant tensor would satisfy . Its four entries acquire factors , , , and , respectively. None equals , so . Equivalently, has no trivial representation. The mixed doublet and third-family terms are again forbidden because their tensor product of group representations contain no trivial representation. Only the product of the two fields survives:
Use the basis of solution for the two doublets, treating the barred left-handed fields as the components to which the printed representation assignment applies. Write and . The doublet contribution to the Yukawa interaction is , so invariant tensors obey
The first equation forces , since their factors are . The second sets . Thus the doublet invariant tensor is . The product of either doublet with has no trivial representation, whereas . Therefore
If instead the unbarred left-handed doublet is chosen to transform by , its barred components transform by ; in that equivalent basis the upper block is . Exchanging the two left components relates these conventions. The singular value decomposition gives the same physical masses in either basis.
No. A counterexample works for every . Take the complete bipartite graph with vertex sets
It is a k-connected graph: deleting fewer than vertices leaves at least one vertex in each side, and the remaining complete bipartite graph is a connected graph.
Choose
and order these sets so that and . Order the other members arbitrarily. These are disjoint sets of cardinality , and every vertex of is one of the specified endpoints.
If the prescribed vertex-disjoint paths existed, no one of them could have an internal vertex, because every vertex must serve as the endpoint of one of the graph paths. Thus each graph path would have to be a single edge. But lie in the same side of the bipartite graph, so there is no edge . This contradicts the requirement on .
For this is the four-vertex cycle graph, with opposite vertices paired. The stronger property that every prescribed pairing can be realized is the linked graph property; vertex connectivity alone does not guarantee it.
Add two new vertices to , joining to every vertex of and to every vertex of , with no other new edges. The vertices are nonadjacent. Suppose has . Since is a k-connected graph, is a connected graph. Both and are nonempty because each original set has cardinality . Choose and ; a graph path between them in extends to an – graph path avoiding .
Consequently no fewer than vertices can form an – vertex separator. By Menger's theorem, there are at least internally vertex-disjoint paths from to . There cannot be more than , since has only neighbours.
Remove from these graph paths. The resulting graph paths in start at distinct vertices of , end at distinct vertices of , and are pairwise vertex-disjoint paths: any shared vertex would have been an internal vertex common to two of the original graph paths. Because there are exactly graph paths, they use every vertex of and as an endpoint. None can pass through another terminal internally, since that terminal already occurs as an endpoint of one of the vertex-disjoint paths. Thus there are vertex-disjoint paths from to , with some pairing of the endpoints. This is also an immediate application of the Set version of Menger theorem.
Use mathematical induction on the maximum degree , including as the base case. A graph of maximum degree zero has no edges, so zero colours suffice.
For , part (ii) supplies a matching in a graph meeting every vertex of vertex degree . Delete its edges, retaining all vertices. Each vertex formerly of maximum degree loses one incident edge; every other vertex already had vertex degree at most . Hence
The resulting graph is still a bipartite graph. By mathematical induction, it has a proper edge colouring using at most colours. Give every edge of one new colour. Since is a matching in a graph, this produces a proper edge colouring of with at most colours. The edges at a vertex of maximum degree require at least colours, so
This is Kőnig's line colouring theorem.
By part (i), choose a matching in a graph meeting all vertices of vertex degree in . Interchanging the two vertex sets gives a matching in a graph meeting all such vertices in . Consider the spanning subgraph with edges .
Every vertex has vertex degree at most two in . An edge belonging to both matchings in a graph forms a whole two-vertex connected component of a graph, since neither endpoint can meet another edge of either matching in a graph. On every other nontrivial connected component of a graph, the edges alternate between and , so the connected component of a graph is a graph path or an even graph cycle. We select one of the two matchings in a graph separately on each connected component of a graph.
On a graph cycle, either choice covers every vertex. On a graph path with an odd number of edges, the first and last edges belong to the same matching in a graph; choose it, and it covers every vertex of the graph path. For a common edge, choose that edge.
It remains to treat a graph path with a positive even number of edges. Its endpoints lie in the same side of the bipartite graph. Every internal vertex meets both matchings in a graph. If both endpoints lie in , choose : every endpoint of vertex degree must already be met by , by its defining property. If both endpoints lie in , choose for the same reason. Thus the choice meets all vertices of vertex degree in that connected component of a graph.
Finally, an isolated vertex of cannot have vertex degree in , since it would have been met by or . The union of the selected edges over the connected components of a graph is a matching in a graph, because different connected components of a graph have disjoint vertex sets. It meets every vertex of vertex degree in . The useful mechanism is the union of two matchings and its alternating connected components of a graph.
Let . Apply Hall's marriage theorem to the induced subgraph on , which is a bipartite graph. For any subset , double counting the edges from to its graph neighbourhood gives
The equality uses the fact that every neighbour of lies in , so all edges incident with have been counted. The final inequality uses the maximum degree bound on . Since , division gives the Hall condition .
Consequently there is a matching in a graph saturating . If is empty, use the empty matching in a graph. This proves the required matching covering maximum-degree vertices in a bipartite graph on the specified side; other vertices need not be met.
Deep currents and tides moving across rough topography, including the Mid-Atlantic Ridge, generate internal waves. Their breaking produces turbulent mixing across isopycnals, called diapycnal mixing. This transfers buoyancy between deep and overlying waters, transforms dense water into lighter water, and contributes to the upward return needed to close the overturning. The required mechanical energy ultimately comes from winds and tides; roughness is a mechanism for converting this input into mixing, rather than an independent source of energy. Mixing is spatially concentrated, and local vertical motion depends on the divergence of the turbulent buoyancy flux, so enhanced mixing does not imply uniform local upwelling everywhere over a ridge.[ref]
Southern Ocean westerly wind stress drives northward surface Ekman transport in the Southern Hemisphere. Its divergence and the circumpolar geometry allow deep water to rise towards the surface along tilted isopycnals, providing an important return route for the deep branch of the Atlantic meridional overturning circulation. Surface heat fluxes and freshwater fluxes then transform the upwelled water before it enters shallower return pathways. The winds supply mechanical energy; the overturning cannot be represented entirely by a local northern density difference. Eddy-induced overturning opposes part of the wind-driven contribution, so the net water-mass transformation and transport need not increase in direct proportion to the wind stress.[ref]

There are unlisted articles, also show them or only show them.