IRC-safe Graph Autoencoder for an unsupervised anomaly detection

Atkinson, Oliver; Bhardwaj, Akanksha; Englert, Christoph; Konar, Partha; Ngairangbam, Vishal S.; Spannowsky, Michael

doi:10.48550/arxiv.2204.12231

Cited by 3 publications

(3 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Inevitably, such searches have led to the use of machine learning tools to sift and find patterns in the data. This field has been attracting an increasing amount of interest from India [327,[872][873][874][875][876] in recent times and holds out the promise of being a major tool in the future runs of the LHC.…”

Section: New Variables For Bsm Studiesmentioning

confidence: 99%

Indian contributions to LHC theory

Raychaudhuri

2023

Eur. Phys. J. Spec. Top.

View full text Add to dashboard Cite

Indian scientists began to work on the theoretical aspects of LHC physics from the early 1980s, at the same time when the rest of the world started taking interest in this then-futuristic topic. From this point grew a whole school of collider phenomenologists, who now form a significant fraction of the Indian high-energy physics community. This article briefly reviews the growth and contributions of the Indian school, on the way describing some of the physics ideas while placing the work in the international context, and proceeding thus, brings the story up to date in mid-2022.Dedicated to the memory of the late D.P. Roy (1941Roy ( -2017, the father of Indian collider theory. The expectant 80sThe LHC owes its origin [1] to CERN's far-sighted leaders, Léon van Hove, who served from 1976-1980 as the Scientific Director-General, and Technical Director-General Sir John Adams (1976)(1977)(1978)(1979)(1980)(1981), who resolved that it would be feasible and scientifically rewarding to build the next-generation storage ring at CERN. Thus, in 1977, Adams wrote the crucial concept note that the LEP collider 1 tunnel, which would be excavated in the 1980s, should be made large enough to accommodate the superconducting magnets required for a future phase of proton acceleration at higher energies. This 1977 note may be thought of as the genesis of the LHC. By 1978, it was already well known in the community, even though the formal approval for the LEP came only in 1981, and the machine itself started running in 1989 [2]. However, the fact that the 20 TeV LHC would eventually come and that the physics at such a machine should be studied, was in the air right from the early 1980s onward. Moreover, the USA was supposed to build a 40 TeV Superconducting Super Collider (SSC) which, sadly, never materialised 2 . The 1980s were, in fact, a period of great optimism for high-energy physics -or, as it was then more commonly known, elementary particle physics. One trigger for this may have lain in the decision of the Nobel Prize committee, in 1979, to recognise the work of Glashow (1961) [3], of Salam (1966-1968 1 The Large Electron-Positron collider.2 It was sanctioned in 1989, and eventually cancelled in 1993.

show abstract

Section: New Variables For Bsm Studiesmentioning

confidence: 99%

Indian contributions to LHC theory

Raychaudhuri

2023

Eur. Phys. J. Spec. Top.

View full text Add to dashboard Cite

show abstract

“…Motivated by their initial success, ML-methods for anomaly detection at the LHC were developed for anomalous jets [7][8][9][10][11][12][13][14][15][16], anomalous events [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35], or to enhance search strategies [36][37][38][39][40][41][42][43][44]. They include a first ATLAS analysis [45], experimental validation [46,47], quantum machine learning [48], self-supervised learning [49,50], applications to heavy-ion collisions [51], the DarkMachines community challenge [52], and the LHC Olympics 2020 community challenge [53,…”

Section: Introductionmentioning

confidence: 99%

A Normalized Autoencoder for LHC Triggers

Dillon¹,

Favaro²,

Plehn³

et al. 2022

Preprint

View full text Add to dashboard Cite

Autoencoders are the ideal analysis tool for the LHC, as they represent its main goal of finding physics beyond the Standard Model. The key challenge is that out-of-distribution anomaly searches based on the compressibility of features do not apply to the LHC, while existing density-based searches lack performance. We present the first autoencoder which identifies anomalous jets symmetrically in the directions of higher and lower complexity. The normalized autoencoder combines a standard bottleneck architecture with a well-defined probabilistic description. It works better than all available autoencoders for top vs QCD jets and reliably identifies different dark-jet signals.

show abstract

“…With the hardware upgrades to FPGAs at the L1 stage and enhancements to the DAQ system at HL-LHC, the integration of unsupervised machine learning algorithms into the L1 trigger system now seems achievable [89][90][91][92]. In this paper, we introduce LLPNet-a lightweight graph autoencoder [93][94][95][96] that utilizes edge convolutions [97] and can detect anomalous signatures of LLPs against the minimum bias and QCD di-jet backgrounds with the restricted tracker level information available at the L1 trigger.…”

mentioning

confidence: 99%

Triggering long-lived particles in HL-LHC and the challenges in the first stage of the trigger system

Bhattacherjee

Mukherjee

Sengupta

et al. 2020

J. High Energ. Phys.

View full text Add to dashboard Cite

Triggering long-lived particles (LLPs) at the first stage of the trigger system is very crucial in LLP searches to ensure that we do not miss them at the very beginning. The future High Luminosity runs of the Large Hadron Collider will have increased number of pileup events per bunch crossing. There will be major upgrades in hardware, firmware and software sides, like tracking at level-1 (L1). The L1 trigger menu will also be modified to cope with pileup and maintain the sensitivity to physics processes. In our study we found that the usual level-1 triggers, mostly meant for triggering prompt particles, will not be very efficient for LLP searches in the 140 pileup environment of HL-LHC, thus pointing to the need to include dedicated L1 triggers in the menu for LLPs. We consider the decay of the LLP into jets and develop dedicated jet triggers using the track information at L1 to select LLP events. We show in our work that these triggers give promising results in identifying LLP events with moderate trigger rates.

show abstract

IRC-safe Graph Autoencoder for an unsupervised anomaly detection

Cited by 3 publications

References 42 publications

Indian contributions to LHC theory

Indian contributions to LHC theory

A Normalized Autoencoder for LHC Triggers

Triggering long-lived particles in HL-LHC and the challenges in the first stage of the trigger system

Contact Info

Product

Resources

About