2.4 Identifying Particle Decays with Jet Substructure
41
HTTv2 [234], introduces a variable jet radius by repeatedly reducing R in the clustering until a significant drop in the jet mass is observed. This value of R has some
additional discrimination power. It is used in a Boosted Decision Tree to improve
the HTT performance, where other input variables are subjet kinematics, ungroomed
and groomed τ N and Qjet volatility (described below).
Mass Jump
The vetoed jet clustering algorithm mass jump [244] is different than other veto
algorithms, as it does not reject pseudojets but prevents them from participating in
the next steps of the jet clustering if this would result in a large increase of the combined jet mass (a mass jump). The algorithm achieves this by a forward clustering,
instead of a decomposition of an existing clustering history. At the start of the algorithm, all pseudojets are labelled active. In the clustering, only active pseudojets are
considered. The condition m i+ j < μ is checked at each step and two pseudojets are
only combined if the condition is met. Else, a mass jump criterion is introduced as
θ m i+ j > max(m i , m j ) and if a mass jump is found the two pseudojets i and j are
labelled as passive and not considered further in the jet clustering. It is also checked
if a mass jump appears between active and passive pseudojets. If no mass jump is
found, the pseudojets i and j are combined and replaced by the resulting pseudojet
in the list of active pseudojets. The clustering terminates once no more active pseudojets are present, where all passive pseudojets are called jets. For θ = 0 and μ = ∞
the algorithm is identical to the standard sequential jet clustering without a veto. As
the vetoed jets are not considered for further clustering, their effective jet radius is
smaller than the parameter R, which now gives an upper bound. The mass jump
algorithm has no inherent grooming, as all jets are kept in the clustering. However,
it is straight forward to reject jets that were labelled passive without fulfilling the
mass jump condition. The algorithm can improve the performance of the HTT, when
employed instead of the MDT declustering for obtaining subjets.
HOTVR Tagger
The Heavy Object Tagger with Variable R (HOTVR) [245] is the only tagger based
on the VR jet algorithm, which adapts R dynamically
11 to the p T of the jet. It includes
a mass jump condition [244, 246] in the clustering process, resulting in subjet finding
and the rejection of soft radiation in one sequence, without the need of declustering
and following grooming steps. A known shortcoming of the VR algorithm is the
clustering of additional radiation into jets in QCD multijet production, resulting in a
higher jet p T on average and an increased rate once a p T selection is applied [187]. The
HOTVR algorithm approaches this issue by modifying the jet clustering procedure
with a veto based on the invariant mass of the pseudojet pair, inspired by mass jump
algorithm [244]. The mass jump veto prevents the recombination of two pseudojets
i and j if the combined invariant mass m i+ j is not large enough. In case a mass
jump is found and p T,i+j > p T,sub the pseudojets are combined, where p T,sub is a
11 The text in this paragraph has been taken from [245] and has been written by the author. It has
been adjusted to fit this book.
41
HTTv2 [234], introduces a variable jet radius by repeatedly reducing R in the clustering until a significant drop in the jet mass is observed. This value of R has some
additional discrimination power. It is used in a Boosted Decision Tree to improve
the HTT performance, where other input variables are subjet kinematics, ungroomed
and groomed τ N and Qjet volatility (described below).
Mass Jump
The vetoed jet clustering algorithm mass jump [244] is different than other veto
algorithms, as it does not reject pseudojets but prevents them from participating in
the next steps of the jet clustering if this would result in a large increase of the combined jet mass (a mass jump). The algorithm achieves this by a forward clustering,
instead of a decomposition of an existing clustering history. At the start of the algorithm, all pseudojets are labelled active. In the clustering, only active pseudojets are
considered. The condition m i+ j < μ is checked at each step and two pseudojets are
only combined if the condition is met. Else, a mass jump criterion is introduced as
θ m i+ j > max(m i , m j ) and if a mass jump is found the two pseudojets i and j are
labelled as passive and not considered further in the jet clustering. It is also checked
if a mass jump appears between active and passive pseudojets. If no mass jump is
found, the pseudojets i and j are combined and replaced by the resulting pseudojet
in the list of active pseudojets. The clustering terminates once no more active pseudojets are present, where all passive pseudojets are called jets. For θ = 0 and μ = ∞
the algorithm is identical to the standard sequential jet clustering without a veto. As
the vetoed jets are not considered for further clustering, their effective jet radius is
smaller than the parameter R, which now gives an upper bound. The mass jump
algorithm has no inherent grooming, as all jets are kept in the clustering. However,
it is straight forward to reject jets that were labelled passive without fulfilling the
mass jump condition. The algorithm can improve the performance of the HTT, when
employed instead of the MDT declustering for obtaining subjets.
HOTVR Tagger
The Heavy Object Tagger with Variable R (HOTVR) [245] is the only tagger based
on the VR jet algorithm, which adapts R dynamically
11 to the p T of the jet. It includes
a mass jump condition [244, 246] in the clustering process, resulting in subjet finding
and the rejection of soft radiation in one sequence, without the need of declustering
and following grooming steps. A known shortcoming of the VR algorithm is the
clustering of additional radiation into jets in QCD multijet production, resulting in a
higher jet p T on average and an increased rate once a p T selection is applied [187]. The
HOTVR algorithm approaches this issue by modifying the jet clustering procedure
with a veto based on the invariant mass of the pseudojet pair, inspired by mass jump
algorithm [244]. The mass jump veto prevents the recombination of two pseudojets
i and j if the combined invariant mass m i+ j is not large enough. In case a mass
jump is found and p T,i+j > p T,sub the pseudojets are combined, where p T,sub is a
11 The text in this paragraph has been taken from [245] and has been written by the author. It has
been adjusted to fit this book.