Machine Learning in High Energy Physics Community White Paper

Albertsson, Kim; Altoe, Piero; Anderson, D.; Anderson, J.; Andrews, Michael Benjamin; Espinosa, Juan Pedro Araque; Aurisano, A.; Basara, L.; Bevan, A. J.; Bhimji, W.; Bonacorsi, D.; Burkle, B.; Calafiura, P.; Campanelli, M.; Capps, Louis; Carminati, F.; Carrazza, Stefano; Chen, Yifan; Childers, J. T.; Coadou, Y.; Coniavitis, E.; Cranmer, K.; David, C.; Davis, D. R.; Simone, Andrea De; Duarte, J.; Erdmann, M.; Eschle, J.; Farbin, A.; Feickert, M.; Castro, N. F.; Fitzpatrick, C.; Floris, M.; Forti, A.; Garra-Tico, J.; Gemmler, J.; Girone, M.; Glaysher, P. C. F.; Gleyzer, S. V.; Gligorov, Vladimir; Golling, T.; Graw, Jonas; Gray, L.; Greenwood, Dick; Hacker, Thomas J.; Harvey, J.; Hegner, B.; Heinrich, L.; Heintz, U.; Hooberman, Ben; Junggeburth, Johannes Josef; Kagan, M.; Kane, Meghan; Kanishchev, K.; Karpiński, Przemysław; Kassabov, Zahari; Kaul, Gautam; Kçira, D.; Keck, Thomas M.; Klimentov, A.; Kowalkowski, Jim; Kreczko, L.; Kurepin, Alexander; Kutschke, Rob; Кузнецов, В. Е.; Köhler, N. M.; Lakomov, I.; Lannon, K.; Lassnig, M.; Limosani, A.; Louppe, Gilles; Mangu, Aashrita; Mató, P.; Narain, Meenakshi; Meinhard, H.; Menasce, D.; Moneta, L.; Moortgat, S.; Neubauer, M. S.; Newman, H. B.; Otten, Sydney; Pabst, Hans; Paganini, M.; Paulini, M.; Perdue, G. N.; Perez, Uzziel; Picazio, A.; Pivarski, J.; Prosper, H.; Psihas, F.; Radovic, A.; Reece, R.; Rinkevicius, A.; Rodrigues, E.; Rorie, J.; Rousseau, D.; Sauers, Aaron; Schramm, S.; Schwartzman, A.; Severini, H.; Seyfert, P.; Siroký, Filip; Skazytkin, Konstantin; Sokoloff, M. D.; Stewart, G. A.; Stienen, Bob; Stockdale, I. E.; Strong, G.; Sun, W.; Thais, S. J.; Tomko, Karen; Upfal, Eli; Usai, E.; Ustyuzhanin, A.; Vala, M.; Vasel, J.; Vallecorsa, S.; Verzetti, M.; Vilasís-Cardona, X.; Vlimant, J. R.; Vukotić, I.; Wang, Sean-Jiun; Watts, G.; Williams, M.; Wu, Wenjing; Wünsch, Stefan; Yang, Kun; Zapata, Óscar

doi:10.48550/arxiv.1807.02876

Cited by 26 publications

(31 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Event selection or categorization based on the information contained in x itself cannot improve the sensitivity of the experiment, and can potentially be detrimental to it. 2 This is illustrated in figure 1, which shows event selection performed based only on x. This only removes some bins from the analysis, hence removing some non-negative, potentially positive, terms from the sum i∈x bins s 2 i /b i (or other measures of sensitivity).…”

Section: Dimensionality Reduction and Complementary Informationmentioning

confidence: 99%

“…Among the more recent additions to the medley of analysis techniques used in experimental HEP are machine learning algorithms. Machine learning techniques have found applications in various aspects of the analysis [1][2][3], and are an active area of research, both from the standpoint of finding new applications as well as refining their current usage.…”

mentioning

confidence: 99%

“…The "signal is better than background" heuristic allows us to rephrase the event selection problem as an event classification problem. Over the last two decades, machine learning methods like artificial neural networks [6] and boosted decision trees [7,8] originally developed to solve classification problems have been successfully employed in collider physics analyses [1][2][3]. These algorithms are trained on simulated data samples to learn features that distinguish signal and background events.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Optimal event selection and categorization in high energy physics, Part 1: Signal discovery

Matchev¹,

Shyamsundar²

2019

Preprint

View full text Add to dashboard Cite

We provide a prescription to train optimal machine-learning-based event selectors and categorizers that maximize the statistical significance of a potential signal excess in high energy physics (HEP) experiments, as quantified by any of six different performance measures. For analyses where the signal search is performed in the distribution of some event variables, our prescription ensures that only the information complementary to those event variables is used in event selection and categorization. This eliminates a major misalignment with the physics goals of the analysis (maximizing the significance of an excess) that exists in the training of typical ML-based event selectors and categorizers. In addition, this decorrelation of event selectors from the relevant event variables prevents the background distribution from becoming peaked in the signal region as a result of event selection, thereby ameliorating the challenges imposed on signal searches by systematic uncertainties. Our event selectors (categorizers) use the output of machine-learning-based classifiers as input and apply optimal selection cutoffs (categorization thresholds) that depend on the event variables being analyzed, as opposed to flat cutoffs (thresholds). These optimal cutoffs and thresholds are learned iteratively, using a novel approach with connections to Lloyd's k-means clustering algorithm. We provide a public, Python 3 implementation of our prescription called ThickBrick, along with usage examples.

show abstract

Section: Dimensionality Reduction and Complementary Informationmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Optimal event selection and categorization in high energy physics, Part 1: Signal discovery

Matchev¹,

Shyamsundar²

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…Machine learning techniques usually perform better with a large amount of data. Data-rich branches of physics like High energy physics and Astronomy thus have employed Machine learning techniques successfully in extracting physical insights from the data [7][8][9][10]. In recent years Machine learning has made its way into other branches like Condensed matter physics and Statistical physics.…”

Section: Introductionmentioning

confidence: 99%

Machine Learned Phase Transitions in a System of Anisotropic Particles on a Square Lattice

Padavala¹,

Singh²,

Kundu³

2021

Preprint

View full text Add to dashboard Cite

The area of Machine learning (ML) has seen exceptional growth in recent years. Successful implementation of ML methods in various branches of physics has led to new insights. These methods have been shown to classify phases in condensed matter systems. Here we study the classification problem of phases in a system of hard rigid rods on a square lattice around a continuous and a discontinuous phase transition using supervised learning (with prior knowledge about the transition points). On comparing a number of ML models we find that convolutional neural network (CNN) classifies the phases with the highest accuracy when only snapshots are given as inputs. We study how the system size affects the model performance. We compare the performance of CNN in classifying the phases around a continuous and a discontinuous phase transition. Further, we show that one can even beat the accuracy of CNN with simpler models by using physics-guided features. Lastly, we show that the critical point in this system can be learned without any prior estimate by using only the information of the ordered phase (as training set). Our study reveals the ML techniques that have been successful in studying spin systems can be easily adapted to more complex systems.

show abstract

“…We utilize both the traditional cut and count search technique as well as machine learning (ML) classifiers to optimize the search efficiency in this channel [39][40][41]. While there exists a plethora of different ML paradigms [42,43], in this work we confine to boosted decision tree approach [44] utilizing the extreme gradient boosting algorithm (XGBoost [45]). Expectedly the ML technique leads to a more aggressive reach in future runs of the LHC at [1.5 − 1.6] TeV for Γ/M in the [0.05 − 0.6] range at 3 ab −1 integrated luminosity.…”

Section: Introductionmentioning

confidence: 99%

Broad toplike vector quarks at LHC and HL-LHC

Dasgupta,

Pramanick,

Ray

2021

Preprint

View full text Add to dashboard Cite

Top like vector quarks arising from underlying strong sectors are expected to have large decay widths pushing them beyond the narrow width approximation. In this paper we consider a broad colored vector quark that strongly couples to an exotic pseudoscalar. We use the full 1PI re-summed propagator for the exotic quark to recast the present LHC constraints ruling out masses below ∼ 1.2 (1.1) TeV for width to mass ratio of 0.1 (0.6). We utilize machine learning techniques that are demonstratively more efficient than traditional cut based searches to present the reach of HL-LHC on the parameter space of this broad resonance. We find that at 3 ab −1 the HL-LHC has a discovery potential up to 1.6 TeV dominated by the pair production channel. We study the feasibility of using machine learning techniques to analyze the broad resonance peaks expected from these exotic quarks at collider experiments like the LHC.

show abstract

Machine Learning in High Energy Physics Community White Paper

Cited by 26 publications

References 42 publications

Optimal event selection and categorization in high energy physics, Part 1: Signal discovery

Optimal event selection and categorization in high energy physics, Part 1: Signal discovery

Machine Learned Phase Transitions in a System of Anisotropic Particles on a Square Lattice

Broad toplike vector quarks at LHC and HL-LHC

Contact Info

Product

Resources

About