diff --git a/HEPML.bib b/HEPML.bib index 53f4f82..422d685 100644 --- a/HEPML.bib +++ b/HEPML.bib @@ -1,5 +1,134 @@ # HEPML Papers +% September 15, 2023 +@article{Ermann:2023unw, + author = "Ermann, Christian and Baker, Stephen and Anber, Mohamed M.", + title = "{Breaking Free with AI: The Deconfinement Transition}", + eprint = "2309.07225", + archivePrefix = "arXiv", + primaryClass = "hep-th", + month = "9", + year = "2023" +} + +% September 14, 2023 +@article{FaucciGiannelli:2023fow, + author = "Faucci Giannelli, Michele and Zhang, Rui", + title = "{CaloShowerGAN, a Generative Adversarial Networks model for fast calorimeter shower simulation}", + eprint = "2309.06515", + archivePrefix = "arXiv", + primaryClass = "physics.ins-det", + month = "9", + year = "2023" +} + +@article{Rusack:2023pob, + author = "Rusack, Roger and Joshi, Bhargav and Alpana, Alpana and Sharma, Seema and Vadnais, Thomas", + title = "{Electron Energy Regression in the CMS High-Granularity Calorimeter Prototype}", + eprint = "2309.06582", + archivePrefix = "arXiv", + primaryClass = "hep-ex", + month = "9", + year = "2023" +} + +@article{Golling:2023mqx, + author = "Golling, Tobias and Klein, Samuel and Mastandrea, Radha and Nachman, Benjamin and Raine, John Andrew", + title = "{Flows for Flows: Morphing one Dataset into another with Maximum Likelihood Estimation}", + eprint = "2309.06472", + archivePrefix = "arXiv", + primaryClass = "hep-ph", + month = "9", + year = "2023" +} + +% September 13, 2023 +@article{Buhmann:2023kdg, + author = {Buhmann, Erik and Gaede, Frank and Kasieczka, Gregor and Korol, Anatolii and Korcari, William and Kr\"uger, Katja and McKeown, Peter}, + title = "{CaloClouds II: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation}", + eprint = "2309.05704", + archivePrefix = "arXiv", + primaryClass = "physics.ins-det", + reportNumber = "DESY-23-130", + month = "9", + year = "2023" +} + +@article{Seong:2023njx, + author = "Seong, Rak-Kyeong", + title = "{Unsupervised Machine Learning Techniques for Exploring Tropical Coamoeba, Brane Tilings and Seiberg Duality}", + eprint = "2309.05702", + archivePrefix = "arXiv", + primaryClass = "hep-th", + reportNumber = "UNIST-MTH-23-RS-04", + month = "9", + year = "2023" +} + +@article{Kashiwa:2023dfx, + author = "Kashiwa, Kouji and Namekawa, Yusuke and Ohnishi, Akira and Takase, Hayato", + title = "{Application of the path optimization method to a discrete spin system}", + eprint = "2309.06018", + archivePrefix = "arXiv", + primaryClass = "hep-lat", + month = "9", + year = "2023" +} + +@article{Badea:2023jdb, + author = "Badea, Anthony and Montejo Berlingen, Javier", + title = "{A data-driven and model-agnostic approach to solving combinatorial assignment problems in searches for new physics}", + eprint = "2309.05728", + archivePrefix = "arXiv", + primaryClass = "hep-ph", + month = "9", + year = "2023" +} + +% September 11, 2023 +@article{Arganda:2023qni, + author = "Arganda, Ernesto and D\'\i{}az, Daniel A. and Perez, Andres D. and Sand\'a Seoane, Rosa M. and Szynkman, Alejandro", + title = "{LHC Study of Third-Generation Scalar Leptoquarks with Machine-Learned Likelihoods}", + eprint = "2309.05407", + archivePrefix = "arXiv", + primaryClass = "hep-ph", + reportNumber = "IFT-UAM/CSIC-23-110", + month = "9", + year = "2023" +} + +% September 07, 2023 +@article{Bedaque:2023udu, + author = "Bedaque, Paulo F. and Kumar, Hersh and Sheng, Andy", + title = "{Neural Network Solutions of Bosonic Quantum Systems in One Dimension}", + eprint = "2309.02352", + archivePrefix = "arXiv", + primaryClass = "nucl-th", + month = "9", + year = "2023" +} + +% September 06, 2023 +@article{Hunt-Smith:2023ccp, + author = "Hunt-Smith, N. T. and Melnitchouk, W. and Ringer, F. and Sato, N. and Thomas, A. W. and White, M. J.", + title = "{Accelerating Markov Chain Monte Carlo sampling with diffusion models}", + eprint = "2309.01454", + archivePrefix = "arXiv", + primaryClass = "hep-ph", + month = "9", + year = "2023" +} + +@article{Sahu:2023uwb, + author = "Sahu, Rameswar and Ghosh, Kirtiman", + title = "{ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower \& Tracker Data Integration}", + eprint = "2309.01568", + archivePrefix = "arXiv", + primaryClass = "hep-ph", + month = "9", + year = "2023" +} + % September 04, 2023 @inproceedings{Detmold:2023kjm, author = "Detmold, William and Kanwar, Gurtej and Lin, Yin and Shanahan, Phiala E. and Wagman, Michael L.", diff --git a/HEPML.tex b/HEPML.tex index 91ecea5..ae56f43 100644 --- a/HEPML.tex +++ b/HEPML.tex @@ -45,7 +45,7 @@ \item Modern reviews~\cite{Larkoski:2017jix,Guest:2018yhq,Albertsson:2018maf,Radovic:2018dip,Carleo:2019ptp,Bourilkov:2019yoi,Schwartz:2021ftp,Karagiorgi:2021ngt,Boehnlein:2021eym,Shanahan:2022ifi} \item Specialized reviews~\cite{Kasieczka:2019dbj,1807719,1808887,Psihas:2020pby,Butter:2020tvl,Forte:2020yip,Brehmer:2020cvb,Nachman:2020ccu,Duarte:2020ngm,Vlimant:2020enz,Cranmer:2019eaq,Rousseau:2020rnz,Kagan:2020yrm,Guan:2020bdl,deLima:2021fwm,Alanazi:2021grv,Baldi:2022okj,Viren:2022qon,Bogatskiy:2022hub,Butter:2022rso,Dvorkin:2022pwo,Adelmann:2022ozp,Thais:2022iok,Harris:2022qtm,Coadou:2022nsh,Benelli:2022sqn,Chen:2022pzc,Plehn:2022ftl,Cheng:2022idp,Huerta:2022kgj,Huber:2022lpm,DeZoort:2023vrm,Du:2023qst,Allaire:2023fgp} \item Classical papers~\cite{Denby:1987rk,Lonnblad:1990bi} - \item Datasets~\cite{Kasieczka:2021xcg,Aarrestad:2021oeb,Benato:2021olt,Govorkova:2021hqu,Chen:2021euv,Qu:2022mxj,Eller:2023myr} + \item Datasets~\cite{Kasieczka:2021xcg,Aarrestad:2021oeb,Benato:2021olt,Govorkova:2021hqu,Chen:2021euv,Qu:2022mxj,Eller:2023myr,Rusack:2023pob} \end{itemize} \item \textbf{Classification} \\\textit{Given a feature space $x\in\mathbb{R}^n$, a binary classifier is a function $f:\mathbb{R}^n\rightarrow [0,1]$, where $0$ corresponds to features that are more characteristic of the zeroth class (e.g. background) and $1$ correspond to features that are more characteristic of the one class (e.g. signal). Typically, $f$ will be a function specified by some parameters $w$ (e.g. weights and biases of a neural network) that are determined by minimizing a loss of the form $L[f]=\sum_{i}\ell(f(x_i),y_i)$, where $y_i\in\{0,1\}$ are labels. The function $\ell$ is smaller when $f(x_i)$ and $y_i$ are closer. Two common loss functions are the mean squared error $\ell(x,y)=(x-y)^2$ and the binary cross entropy $\ell(x,y)=y\log(x)+(1-y)\log(1-x)$. Exactly what `more characteristic of' means depends on the loss function used to determine $f$. It is also possible to make a multi-class classifier. A common strategy for the multi-class case is to represent each class as a different basis vector in $\mathbb{R}^{n_\text{classes}}$ and then $f(x)\in[0,1]^{n_\text{classes}}$. In this case, $f(x)$ is usually restricted to have its $n_\text{classes}$ components sum to one and the loss function is typically the cross entropy $\ell(x,y)=\sum_\text{classes $i$} y_i\log(x)$.} @@ -65,7 +65,7 @@ \\\textit{Recursive neural networks are natural tools for processing data in a tree structure.} \item \textbf{Graphs}~\cite{Henrion:DLPS2017,Ju:2020xty,Abdughani:2018wrw,Martinez:2018fwc,Ren:2019xhp,Moreno:2019bmu,Qasim:2019otl,Chakraborty:2019imr,DiBello:2020bas,Chakraborty:2020yfc,1797439,1801423,1808887,Iiyama:2020wap,1811770,Choma:2020cry,alonsomonsalve2020graph,guo2020boosted,Heintz:2020soy,Verma:2020gnq,Dreyer:2020brq,Qian:2021vnh,Pata:2021oez,Biscarat:2021dlj,Rossi:2021tjf,Hewes:2021heg,Thais:2021qcb,Dezoort:2021kfk,Verma:2021ceh,Hariri:2021clz,Belavin:2021bxb,Atkinson:2021nlt,Konar:2021zdg,Atkinson:2021jnj,Tsan:2021brw,Elabd:2021lgo,Pata:2022wam,Gong:2022lye,Qasim:2022rww,Ma:2022bvt,Bogatskiy:2022czk,Builtjes:2022usj,DiBello:2022iwf,Mokhtar:2022pwm,Huang:2023ssr,Forestano:2023fpj,Anisha:2023xmh,Ehrke:2023cpn,Murnane:2023kfm,Yu:2023juh,Neu:2023sfh,Wang:2023cac,McEneaney:2023vwp,Liu:2023siw,GarciaPardinas:2023pmx,Duperrin:2023elp,BelleII:2023egc,Holmberg:2023rfr,Bhattacherjee:2023evs,Murnane:2023ksa} \\\textit{A graph is a collection of nodes and edges. Graph neural networks are natural tools for processing data in a tree structure.} - \item \textbf{Sets (point clouds)}~\cite{Komiske:2018cqr,Qu:2019gqs,Mikuni:2020wpr,Shlomi:2020ufi,Dolan:2020qkr,Fenton:2020woz,Lee:2020qil,collado2021learning,Mikuni:2021pou,Shmakov:2021qdz,Shimmin:2021pkm,ATL-PHYS-PUB-2020-014,Qu:2022mxj,Kach:2022uzq,Onyisi:2022hdh,Athanasakos:2023fhq,Kach:2023rqw} + \item \textbf{Sets (point clouds)}~\cite{Komiske:2018cqr,Qu:2019gqs,Mikuni:2020wpr,Shlomi:2020ufi,Dolan:2020qkr,Fenton:2020woz,Lee:2020qil,collado2021learning,Mikuni:2021pou,Shmakov:2021qdz,Shimmin:2021pkm,ATL-PHYS-PUB-2020-014,Qu:2022mxj,Kach:2022uzq,Onyisi:2022hdh,Athanasakos:2023fhq,Kach:2023rqw,Badea:2023jdb} \\\textit{A point cloud is a (potentially variable-size) set of points in space. Sets are distinguished from sequences in that there is no particular order (i.e. permutation invariance). Sets can also be viewed as graphs without edges and so graph methods that can parse variable-length inputs may also be appropriate for set learning, although there are other methods as well.} \item \textbf{Physics-inspired basis}~\cite{Datta:2019,Datta:2017rhs,Datta:2017lxt,Komiske:2017aww,Butter:2017cot,Grojean:2020ech,Kishimoto:2022eum,Larkoski:2023nye,Munoz:2023csn} \\\textit{This is a catch-all category for learning using other representations that use some sort of manual or automated physics-preprocessing.} @@ -78,7 +78,7 @@ \\\textit{Due to the fidelity of $b$-tagging, boosted, hadronically decaying Higgs bosons (predominantly decaying to $b\bar{b}$) has unique challenged and opportunities compared with $W/Z$ tagging.} \item \textbf{quarks and gluons}~\cite{ATL-PHYS-PUB-2017-017,Komiske:2016rsd,Cheng:2017rdo,Stoye:DLPS2017,Chien:2018dfn,Moreno:2019bmu,Kasieczka:2018lwf,1806025,Lee:2019ssx,Lee:2019cad,Dreyer:2020brq,Romero:2021qlf,Filipek:2021qbe,Dreyer:2021hhr,Bright-Thonney:2022xkx,CrispimRomao:2023ssj,Athanasakos:2023fhq,He:2023cfc,Shen:2023ofd} \\\textit{Quark jets tend to be narrower and have fewer particles than gluon jets. This classification task has been a benchmark for many new machine learning models.} - \item \textbf{top quark} tagging~\cite{Almeida:2015jua,Stoye:DLPS2017,Kasieczka:2019dbj,Chakraborty:2020yfc,Diefenbacher:2019ezd,Butter:2017cot,Kasieczka:2017nvn,Macaluso:2018tck,Bhattacharya:2020vzu,Lim:2020igi,Dreyer:2020brq,Aguilar-Saavedra:2021rjk,Andrews:2021ejw,Dreyer:2022yom,Ahmed:2022hct,Munoz:2022gjq,Bhattacherjee:2022gjq,Choi:2023slq,Keicher:2023mer,He:2023cfc,Bogatskiy:2023nnw,Shen:2023ofd,Isildak:2023dnf} + \item \textbf{top quark} tagging~\cite{Almeida:2015jua,Stoye:DLPS2017,Kasieczka:2019dbj,Chakraborty:2020yfc,Diefenbacher:2019ezd,Butter:2017cot,Kasieczka:2017nvn,Macaluso:2018tck,Bhattacharya:2020vzu,Lim:2020igi,Dreyer:2020brq,Aguilar-Saavedra:2021rjk,Andrews:2021ejw,Dreyer:2022yom,Ahmed:2022hct,Munoz:2022gjq,Bhattacherjee:2022gjq,Choi:2023slq,Keicher:2023mer,He:2023cfc,Bogatskiy:2023nnw,Shen:2023ofd,Isildak:2023dnf,Sahu:2023uwb} \\\textit{Boosted top quarks form jets that have a three-prong substructure ($t\rightarrow Wb,W\rightarrow q\bar{q}$).} \item \textbf{strange jets}~\cite{Nakai:2020kuu,Erdmann:2019blf,Erdmann:2020ovh,Subba:2023rpm} \\\textit{Strange quarks have a very similar fragmentation to generic quark and gluon jets, so this is a particularly challenging task.} @@ -98,7 +98,7 @@ \\\textit{Machine learning is often used in astrophysics and cosmology in different ways than terrestrial particle physics experiments due to a general divide between Bayesian and Frequentist statistics. However, there are many similar tasks and a growing number of proposals designed for one domain that apply to the other. See also https://github.com/georgestein/ml-in-cosmology.} \item \textbf{Tracking}~\cite{Farrell:DLPS2017,Farrell:2018cjr,Amrouche:2019wmx,Ju:2020xty,Akar:2020jti,Shlomi:2020ufi,Choma:2020cry,Siviero:2020tim,Fox:2020hfm,Amrouche:2021tlm,goto2021development,Biscarat:2021dlj,Akar:2021gns,Thais:2021qcb,Ju:2021ayy,Dezoort:2021kfk,Edmonds:2021lzd,Lavrik:2021zgt,Huth:2021zcm,Goncharov:2021wvd,Wang:2022oer,Alonso-Monsalve:2022zlm,Bakina:2022mhs,Akram:2022zmj,Sun:2022bxx,Abidi:2022ogh,Bae:2023eec,Knipfer:2023zrv,Akar:2023zhd} \\\textit{Charged particle tracking is a challenging pattern recognition task. This category is for various classification tasks associated with tracking, such as seed selection.} - \item \textbf{Heavy Ions / Nuclear Physics}~\cite{Pang:2016vdc,Chien:2018dfn,Du:2020pmp,Du:2019civ,Mallick:2021wop,Nagu:2021zho,Zhao:2021yjo,Sombillo:2021ifs,Zhou:2021bvw,Apolinario:2021olp,Brown:2021upr,Du:2021pqa,Kuttan:2021npg,Huang:2021iux,Shokr:2021ouh,He:2021uko,Habashy:2021orz,Zepeda:2021tzp,Mishra:2021eqb,Ng:2021ibr,Habashy:2021qku,Biro:2021zgm,Lai:2021ckt,Du:2021qwv,Du:2021brx,Xiang:2021ssj,Soma:2022qnv,Rahman:2022tfq,Boglione:2022gpv,Liyanage:2022byj,Liu:2022hzd,Fanelli:2022kro,Chen:2022shj,Saha:2022skj,Lee:2022kdn,Biro:2022zhl,Zhang:2022hjh,Yang:2022eag,Rigo:2022ces,Yang:2022rlw,Munoz:2022slm,Goriely:2022upe,Mallick:2022alr,Fore:2022ljl,Steffanic:2023cyx,Mallick:2023vgi,He:2023urp,Xu:2023fbs,Kanwar:2023otc,Mumpower:2023lch,Escher:2023oyy,Hirvonen:2023lqy,Biro:2023kyx,He:2023zin,Zhou:2023pti,CrispimRomao:2023ssj,Basak:2023wzq,Shi:2023xfz,Soleymaninia:2023dds,Lin:2023bmy,Dellen:2023avd,AlHammal:2023svo,Wang:2023muv,Wang:2023kcg,Ai:2023azx,Yiu:2023ido,Karmakar:2023mhy,Lasseri:2023dhi,Yoshida:2023wrb,Liu:2023xgl,Hizawa:2023plv,Wen:2023oju,Allaire:2023fgp} + \item \textbf{Heavy Ions / Nuclear Physics}~\cite{Pang:2016vdc,Chien:2018dfn,Du:2020pmp,Du:2019civ,Mallick:2021wop,Nagu:2021zho,Zhao:2021yjo,Sombillo:2021ifs,Zhou:2021bvw,Apolinario:2021olp,Brown:2021upr,Du:2021pqa,Kuttan:2021npg,Huang:2021iux,Shokr:2021ouh,He:2021uko,Habashy:2021orz,Zepeda:2021tzp,Mishra:2021eqb,Ng:2021ibr,Habashy:2021qku,Biro:2021zgm,Lai:2021ckt,Du:2021qwv,Du:2021brx,Xiang:2021ssj,Soma:2022qnv,Rahman:2022tfq,Boglione:2022gpv,Liyanage:2022byj,Liu:2022hzd,Fanelli:2022kro,Chen:2022shj,Saha:2022skj,Lee:2022kdn,Biro:2022zhl,Zhang:2022hjh,Yang:2022eag,Rigo:2022ces,Yang:2022rlw,Munoz:2022slm,Goriely:2022upe,Mallick:2022alr,Fore:2022ljl,Steffanic:2023cyx,Mallick:2023vgi,He:2023urp,Xu:2023fbs,Kanwar:2023otc,Mumpower:2023lch,Escher:2023oyy,Hirvonen:2023lqy,Biro:2023kyx,He:2023zin,Zhou:2023pti,CrispimRomao:2023ssj,Basak:2023wzq,Shi:2023xfz,Soleymaninia:2023dds,Lin:2023bmy,Dellen:2023avd,AlHammal:2023svo,Wang:2023muv,Wang:2023kcg,Ai:2023azx,Yiu:2023ido,Karmakar:2023mhy,Lasseri:2023dhi,Yoshida:2023wrb,Liu:2023xgl,Hizawa:2023plv,Wen:2023oju,Allaire:2023fgp,Bedaque:2023udu} \\\textit{Many tools in high energy nuclear physics are similar to high energy particle physics. The physics target of these studies are to understand collective properties of the strong force.} \end{itemize} \item \textbf{Learning strategies} @@ -108,7 +108,7 @@ \\\textit{In addition to learnable weights $w$, classifiers have a number of non-differentiable parameters like the number of layers in a neural network. These parameters are called hyperparameters.} \item \textbf{Weak/Semi supervision}~\cite{Dery:2017fap,Metodiev:2017vrx,Komiske:2018oaa,Collins:2018epr,Collins:2019jip,Borisyak:2019vbz,Cohen:2017exh,Komiske:2018vkc,Metodiev:2018ftz,collaboration2020dijet,Amram:2020ykb,Brewer:2020och,Dahbi:2020zjw,Lee:2019ssx,Lieberman:2021krq,Komiske:2022vxg,Li:2022omf,Finke:2022lsu,LeBlanc:2022bwd,Dolan:2022ikg,Bardhan:2023mia,Witkowski:2023htt} \\\textit{For supervised learning, the labels $y_i$ are known. In the case that the labels are noisy or only known with some uncertainty, then the learning is called weak supervision. Semi-supervised learning is the related case where labels are known for only a fraction of the training examples.} - \item \textbf{Unsupervised}~\cite{Mackey:2015hwa,Komiske:2019fks,1797846,Dillon:2019cqt,Cai:2020vzx,Howard:2021pos,Dillon:2021gag,Huang:2023kgs,Kitouni:2023rct} + \item \textbf{Unsupervised}~\cite{Mackey:2015hwa,Komiske:2019fks,1797846,Dillon:2019cqt,Cai:2020vzx,Howard:2021pos,Dillon:2021gag,Huang:2023kgs,Kitouni:2023rct,Badea:2023jdb} \\\textit{When no labels are provided, the learning is called unsupervised.} \item \textbf{Reinforcement Learning}~\cite{Carrazza:2019efs,Brehmer:2020brs,John:2020sak,Harvey:2021oue,Cranmer:2021gdt,Windisch:2021mem,Dersy:2022bym,Nishimura:2023wdu} \\\textit{Instead of learning to distinguish different types of examples, the goal of reinforcement learning is to learn a strategy (policy). The prototypical example of reinforcement learning in learning a strategy to play video games using some kind of score as a feedback during the learning.} @@ -149,7 +149,7 @@ \\\textit{The target features could be parameters of a model, which can be learned directly through a regression setup. Other forms of inference are described in later sections (which could also be viewed as regression).} \item \textbf{Parton Distribution Functions (and related)}~\cite{DelDebbio:2020rgv,Grigsby:2020auv,Rossi:2020sbh,Carrazza:2021hny,Ball:2021leu,Ball:2021xlu,Khalek:2021gon,Iranipour:2022iak,Gao:2022uhg,Gao:2022srd,Candido:2023utz,Wang:2023nab,Kassabov:2023hbm,Wang:2023poi,Fernando:2023obn,Rabemananjara:2023xfq} \\\textit{Various machine learning models can provide flexible function approximators, which can be useful for modeling functions that cannot be determined easily from first principles such as parton distribution functions.} - \item \textbf{Lattice Gauge Theory}~\cite{Kanwar:2003.06413,Favoni:2020reg,Bulusu:2021rqz,Shi:2021qri,Hackett:2021idh,Yoon:2018krb,Zhang:2019qiq,Nguyen:2019gpo,Favoni:2021epq,Chen:2021jey,Bulusu:2021njs,Shi:2022yqw,Luo:2022jzl,Chen:2022ytr,Li:2022ozl,Kang:2022jbg,Albandea:2022fky,Khan:2022vot,Sale:2022snt,Kim:2022rna,Karsch:2022yka,Favoni:2022mcg,Chen:2022asj,Bacchio:2022vje,Bacchio:2022vje,Gao:2022uhg,Aguilar:2022thg,Lawrence:2022dba,Peng:2022wdl,Lehner:2023bba,Albandea:2023wgd,Nicoli:2023qsl,Aronsson:2023rli,Zhou:2023pti,Hudspith:2023loy,R:2023dcr,Bender:2023gwr,NarcisoFerreira:2023kak,Lehner:2023prf,Singha:2023xxq,Riberdy:2023awf,Buzzicotti:2023qdv,Caselle:2023mvh,Detmold:2023kjm} + \item \textbf{Lattice Gauge Theory}~\cite{Kanwar:2003.06413,Favoni:2020reg,Bulusu:2021rqz,Shi:2021qri,Hackett:2021idh,Yoon:2018krb,Zhang:2019qiq,Nguyen:2019gpo,Favoni:2021epq,Chen:2021jey,Bulusu:2021njs,Shi:2022yqw,Luo:2022jzl,Chen:2022ytr,Li:2022ozl,Kang:2022jbg,Albandea:2022fky,Khan:2022vot,Sale:2022snt,Kim:2022rna,Karsch:2022yka,Favoni:2022mcg,Chen:2022asj,Bacchio:2022vje,Bacchio:2022vje,Gao:2022uhg,Aguilar:2022thg,Lawrence:2022dba,Peng:2022wdl,Lehner:2023bba,Albandea:2023wgd,Nicoli:2023qsl,Aronsson:2023rli,Zhou:2023pti,Hudspith:2023loy,R:2023dcr,Bender:2023gwr,NarcisoFerreira:2023kak,Lehner:2023prf,Singha:2023xxq,Riberdy:2023awf,Buzzicotti:2023qdv,Caselle:2023mvh,Detmold:2023kjm,Kashiwa:2023dfx,Ermann:2023unw} \\\textit{Lattice methods offer a complementary approach to perturbation theory. A key challenge is to create approaches that respect the local gauge symmetry (equivariant networks).} \item \textbf{Function Approximation}~\cite{1853982,Haddadin:2021mmo,Chahrour:2021eiv,Wang:2021jou,Kitouni:2021fkh,Lei:2022dvn,Wang:2023nab,Fernando:2023obn} \\\textit{Approximating functions that obey certain (physical) constraints.} @@ -165,13 +165,13 @@ \item \textbf{Generative models / density estimation} \\\textit{The goal of generative modeling is to learn (explicitly or implicitly) a probability density $p(x)$ for the features $x\in\mathbb{R}^n$. This task is usually unsupervised (no labels).} \begin{itemize} - \item \textbf{GANs}~\cite{deOliveira:2017pjk,Paganini:2017hrr,Paganini:2017dwg,Alonso-Monsalve:2018aqs,Butter:2019eyo,Martinez:2019jlu,Bellagente:2019uyp,Vallecorsa:2019ked,SHiP:2019gcl,Carrazza:2019cnt,Butter:2019cae,Lin:2019htn,DiSipio:2019imz,Hashemi:2019fkn,Chekalina:2018hxi,ATL-SOFT-PUB-2018-001,Zhou:2018ill,Carminati:2018khv,Vallecorsa:2018zco,Datta:2018mwd,Musella:2018rdi,Erdmann:2018kuh,Deja:2019vcv,Derkach:2019qfk,Erbin:2018csv,Erdmann:2018jxd,Urban:2018tqv,Oliveira:DLPS2017,deOliveira:2017rwa,Farrell:2019fsm,Hooberman:DLPS2017,Belayneh:2019vyx,Wang:2020tap,buhmann2020getting,Alanazi:2020jod,2009.03796,2008.06545,Kansal:2020svm,Maevskiy:2020ank,Lai:2020byl,Choi:2021sku,Rehm:2021zow,Rehm:2021zoz,Carrazza:2021hny,Rehm:2021qwm,Lebese:2021foi,Winterhalder:2021ave,Kansal:2021cqp,NEURIPS2020_a878dbeb,Khattak:2021ndw,Mu:2021nno,Li:2021cbp,Bravo-Prieto:2021ehz,Anderlini:2021qpm,Chisholm:2021pdn,Desai:2021wbb,Buhmann:2021caf,Bieringer:2022cbs,Ghosh:2022zdz,Anderlini:2022ckd,Ratnikov:2022hge,Rogachev:2022hjg,ATLAS:2022jhk,Anderlini:2022hgm,Buhmann:2023pmh,Yue:2023uva,Hashemi:2023ruu,EXO:2023pkl,Diefenbacher:2023prl,Chan:2023ume,Dubinski:2023fsy,Alghamdi:2023emm,Barbetti:2023bvi,Erdmann:2023ngr} + \item \textbf{GANs}~\cite{deOliveira:2017pjk,Paganini:2017hrr,Paganini:2017dwg,Alonso-Monsalve:2018aqs,Butter:2019eyo,Martinez:2019jlu,Bellagente:2019uyp,Vallecorsa:2019ked,SHiP:2019gcl,Carrazza:2019cnt,Butter:2019cae,Lin:2019htn,DiSipio:2019imz,Hashemi:2019fkn,Chekalina:2018hxi,ATL-SOFT-PUB-2018-001,Zhou:2018ill,Carminati:2018khv,Vallecorsa:2018zco,Datta:2018mwd,Musella:2018rdi,Erdmann:2018kuh,Deja:2019vcv,Derkach:2019qfk,Erbin:2018csv,Erdmann:2018jxd,Urban:2018tqv,Oliveira:DLPS2017,deOliveira:2017rwa,Farrell:2019fsm,Hooberman:DLPS2017,Belayneh:2019vyx,Wang:2020tap,buhmann2020getting,Alanazi:2020jod,2009.03796,2008.06545,Kansal:2020svm,Maevskiy:2020ank,Lai:2020byl,Choi:2021sku,Rehm:2021zow,Rehm:2021zoz,Carrazza:2021hny,Rehm:2021qwm,Lebese:2021foi,Winterhalder:2021ave,Kansal:2021cqp,NEURIPS2020_a878dbeb,Khattak:2021ndw,Mu:2021nno,Li:2021cbp,Bravo-Prieto:2021ehz,Anderlini:2021qpm,Chisholm:2021pdn,Desai:2021wbb,Buhmann:2021caf,Bieringer:2022cbs,Ghosh:2022zdz,Anderlini:2022ckd,Ratnikov:2022hge,Rogachev:2022hjg,ATLAS:2022jhk,Anderlini:2022hgm,Buhmann:2023pmh,Yue:2023uva,Hashemi:2023ruu,EXO:2023pkl,Diefenbacher:2023prl,Chan:2023ume,Dubinski:2023fsy,Alghamdi:2023emm,Barbetti:2023bvi,Erdmann:2023ngr,FaucciGiannelli:2023fow} \\\textit{Generative Adversarial Networks~\cite{Goodfellow:2014upx} learn $p(x)$ implicitly through the minimax optimization of two networks: one that maps noise to structure $G(z)$ and one a classifier (called the discriminator) that learns to distinguish examples generated from $G(z)$ and those generated from the target process. When the discriminator is maximally `confused', then the generator is effectively mimicking $p(x)$.} \item \textbf{Autoencoders}~\cite{Monk:2018zsb,ATL-SOFT-PUB-2018-001,Cheng:2020dal,1816035,Howard:2021pos,Buhmann:2021lxj,Bortolato:2021zic,deja2020endtoend,Hariri:2021clz,Fanelli:2019qaq,Collins:2021pld,Orzari:2021suh,Jawahar:2021vyu,Tsan:2021brw,Buhmann:2021caf,Touranakou:2022qrp,Ilten:2022jfm,Collins:2022qpr,AbhishekAbhishek:2022wby,Cresswell:2022tof,Roche:2023int,Anzalone:2023ugq,Lasseri:2023dhi,Chekanov:2023uot} \\\textit{An autoencoder consists of two functions: one that maps $x$ into a latent space $z$ (encoder) and a second one that maps the latent space back into the original space (decoder). The encoder and decoder are simultaneously trained so that their composition is nearly the identity. When the latent space has a well-defined probability density (as in variational autoencoders), then one can sample from the autoencoder by applying the detector to a randomly chosen element of the latent space.} - \item \textbf{Normalizing flows}~\cite{Albergo:2019eim,1800956,Kanwar:2003.06413,Brehmer:2020vwc,Bothmann:2020ywa,Gao:2020zvv,Gao:2020vdv,Nachman:2020lpy,Choi:2020bnf,Lu:2020npg,Bieringer:2020tnw,Hollingsworth:2021sii,Winterhalder:2021ave,Krause:2021ilc,Hackett:2021idh,Menary:2021tjg,Hallin:2021wme,NEURIPS2020_a878dbeb,Vandegar:2020yvw,Jawahar:2021vyu,Bister:2021arb,Krause:2021wez,Butter:2021csz,Winterhalder:2021ngy,Butter:2022lkf,Verheyen:2022tov,Leigh:2022lpn,Chen:2022ytr,Albandea:2022fky,Krause:2022jna,Cresswell:2022tof,Kach:2022qnf,Kach:2022uzq,Dolan:2022ikg,Backes:2022vmn,Heimel:2022wyj,Albandea:2023wgd,Rousselot:2023pcj,Diefenbacher:2023vsw,Nicoli:2023qsl,R:2023dcr,Nachman:2023clf,Raine:2023fko,Golling:2023yjq,Wen:2023oju,Xu:2023xdc,Singha:2023xxq,Buckley:2023rez,Pang:2023wfx} + \item \textbf{Normalizing flows}~\cite{Albergo:2019eim,1800956,Kanwar:2003.06413,Brehmer:2020vwc,Bothmann:2020ywa,Gao:2020zvv,Gao:2020vdv,Nachman:2020lpy,Choi:2020bnf,Lu:2020npg,Bieringer:2020tnw,Hollingsworth:2021sii,Winterhalder:2021ave,Krause:2021ilc,Hackett:2021idh,Menary:2021tjg,Hallin:2021wme,NEURIPS2020_a878dbeb,Vandegar:2020yvw,Jawahar:2021vyu,Bister:2021arb,Krause:2021wez,Butter:2021csz,Winterhalder:2021ngy,Butter:2022lkf,Verheyen:2022tov,Leigh:2022lpn,Chen:2022ytr,Albandea:2022fky,Krause:2022jna,Cresswell:2022tof,Kach:2022qnf,Kach:2022uzq,Dolan:2022ikg,Backes:2022vmn,Heimel:2022wyj,Albandea:2023wgd,Rousselot:2023pcj,Diefenbacher:2023vsw,Nicoli:2023qsl,R:2023dcr,Nachman:2023clf,Raine:2023fko,Golling:2023yjq,Wen:2023oju,Xu:2023xdc,Singha:2023xxq,Buckley:2023rez,Pang:2023wfx,Golling:2023mqx} \\\textit{Normalizing flows~\cite{pmlr-v37-rezende15} learn $p(x)$ explicitly by starting with a simple probability density and then applying a series of bijective transformations with tractable Jacobians.} - \item \textbf{Diffusion Models}~\cite{Mikuni:2022xry,Leigh:2023toe,Mikuni:2023dvk,Shmakov:2023kjj,Buhmann:2023bwk,Butter:2023fov,Mikuni:2023tok,Acosta:2023zik,Imani:2023blb,Amram:2023onf,Diefenbacher:2023flw,Cotler:2023lem,Diefenbacher:2023wec,Mikuni:2023tqg} + \item \textbf{Diffusion Models}~\cite{Mikuni:2022xry,Leigh:2023toe,Mikuni:2023dvk,Shmakov:2023kjj,Buhmann:2023bwk,Butter:2023fov,Mikuni:2023tok,Acosta:2023zik,Imani:2023blb,Amram:2023onf,Diefenbacher:2023flw,Cotler:2023lem,Diefenbacher:2023wec,Mikuni:2023tqg,Hunt-Smith:2023ccp,Buhmann:2023kdg} \\\textit{These approaches learn the gradient of the density instead of the density directly.} \item \textbf{Transformer Models}~\cite{Finke:2023veq,Butter:2023fov,Raine:2023fko} \\\textit{These approaches learn the density or perform generative modeling using transformer-based networks.} @@ -183,7 +183,7 @@ \\\textit{Monte Carlo event generators integrate over a phase space that needs to be generated efficiently and this can be aided by machine learning methods.} \item \textbf{Gaussian processes}~\cite{Frate:2017mai,Bertone:2016mdy,1804325,Cisbani:2019xta} \\\textit{These are non-parametric tools for modeling the `time'-dependence of a random variable. The `time' need not be actual time - for instance, one can use Gaussian processes to model the energy dependence of some probability density.} - \item \textbf{Other/hybrid}~\cite{Cresswell:2022tof,DiBello:2022rss,Li:2022jon,Kansal:2022spb,Butter:2023fov,Kronheim:2023jrl,Santos:2023mib} + \item \textbf{Other/hybrid}~\cite{Cresswell:2022tof,DiBello:2022rss,Li:2022jon,Kansal:2022spb,Butter:2023fov,Kronheim:2023jrl,Santos:2023mib,Sahu:2023uwb} \\\textit{Architectures that combine different network elements or otherwise do not fit into the other categories.} \end{itemize} \item \textbf{Anomaly detection}~\cite{DAgnolo:2018cun,Collins:2018epr,Collins:2019jip,DAgnolo:2019vbw,Farina:2018fyg,Heimel:2018mkt,Roy:2019jae,Cerri:2018anq,Blance:2019ibf,Hajer:2018kqm,DeSimone:2018efk,Mullin:2019mmh,1809.02977,Dillon:2019cqt,Andreassen:2020nkr,Nachman:2020lpy,Aguilar-Saavedra:2017rzt,Romao:2019dvs,Romao:2020ojy,knapp2020adversarially,collaboration2020dijet,1797846,1800445,Amram:2020ykb,Cheng:2020dal,Khosa:2020qrz,Thaprasop:2020mzp,Alexander:2020mbx,aguilarsaavedra2020mass,1815227,pol2020anomaly,Mikuni:2020qds,vanBeekveld:2020txa,Park:2020pak,Faroughy:2020gas,Stein:2020rou,Kasieczka:2021xcg,Chakravarti:2021svb,Batson:2021agz,Blance:2021gcs,Bortolato:2021zic,Collins:2021nxn,Dillon:2021nxw,Finke:2021sdf,Shih:2021kbt,Atkinson:2021nlt,Kahn:2021drv,Aarrestad:2021oeb,Dorigo:2021iyy,Caron:2021wmq,Govorkova:2021hqu,Kasieczka:2021tew,Volkovich:2021txe,Govorkova:2021utb,Hallin:2021wme,Ostdiek:2021bem,Fraser:2021lxm,Jawahar:2021vyu,Herrero-Garcia:2021goa,Aguilar-Saavedra:2021utu,Tombs:2021wae,Lester:2021aks,Mikuni:2021nwn,Chekanov:2021pus,dAgnolo:2021aun,Canelli:2021aps,Ngairangbam:2021yma,Bradshaw:2022qev,Aguilar-Saavedra:2022ejy,Buss:2022lxw,Alvi:2022fkk,Dillon:2022tmm,Birman:2022xzu,Raine:2022hht,Letizia:2022xbe,Fanelli:2022xwl,Finke:2022lsu,Verheyen:2022tov,Dillon:2022mkq,Caron:2022wrw,Park:2022zov,Kamenik:2022qxs,Hallin:2022eoq,Kasieczka:2022naq,Araz:2022zxk,Mastandrea:2022vas,Schuhmacher:2023pro,Roche:2023int,Golling:2023juz,Sengupta:2023xqy,Mikuni:2023tok,Golling:2023yjq,Vaslin:2023lig,ATLAS:2023azi,Chekanov:2023uot} @@ -197,7 +197,7 @@ \\\textit{This is the task of removing detector distortions. In contrast to parameter estimation, the goal is not to infer model parameters, but instead, the undistorted phase space probability density. This is often also called deconvolution.} \item \textbf{Domain adaptation}~\cite{Rogozhnikov:2016bdp,Andreassen:2019nnm,Cranmer:2015bka,2009.03796,Nachman:2021opi,Camaiani:2022kul,Schreck:2023pzs,Algren:2023qnb} \\\textit{Morphing simulations to look like data is a form of domain adaptation.} - \item \textbf{BSM}~\cite{Andreassen:2020nkr,Hollingsworth:2020kjg,Brehmer:2018kdj,Brehmer:2018eca,Brehmer:2018hga,Brehmer:2019xox,Romao:2020ojy,deSouza:2022uhk,GomezAmbrosio:2022mpm,Castro:2022zpq,Anisha:2023xmh,Dennis:2023kfe,vanBeekveld:2023ney,Chhibra:2023tyf,Mandal:2023mck,Franz:2023gic} + \item \textbf{BSM}~\cite{Andreassen:2020nkr,Hollingsworth:2020kjg,Brehmer:2018kdj,Brehmer:2018eca,Brehmer:2018hga,Brehmer:2019xox,Romao:2020ojy,deSouza:2022uhk,GomezAmbrosio:2022mpm,Castro:2022zpq,Anisha:2023xmh,Dennis:2023kfe,vanBeekveld:2023ney,Chhibra:2023tyf,Mandal:2023mck,Franz:2023gic,Arganda:2023qni} \\\textit{This category is for parameter estimation when the parameter is the signal strength of new physics.} \item \textbf{Differentiable Simulation}~\cite{Heinrich:2022xfa,Nachman:2022jbj,Lei:2022dvn,Napolitano:2023jhg,Shenoy:2023ros,Kagan:2023gxz} \\\textit{Coding up a simulation using a differentiable programming language like TensorFlow, PyTorch, or JAX.} @@ -218,7 +218,7 @@ \\\textit{ML can also be utilized in formal theory.} \begin{itemize} \item Theory and physics for ML~\cite{Erbin:2022lls,Zuniga-Galindo:2023hty,Banta:2023kqe,Zuniga-Galindo:2023uwp,Kumar:2023hlu,Demirtas:2023fir} - \item ML for theory~\cite{Berglund:2022gvm,Erbin:2022rgx,Gerdes:2022nzr,Escalante-Notario:2022fik,Chen:2022jwd,Cheung:2022itk,He:2023csq,Lal:2023dkj,Dorrill:2023vox,Forestano:2023ijh,Dersy:2023job,Cotler:2023lem,Mizera:2023bsw,Gnech:2023prs} + \item ML for theory~\cite{Berglund:2022gvm,Erbin:2022rgx,Gerdes:2022nzr,Escalante-Notario:2022fik,Chen:2022jwd,Cheung:2022itk,He:2023csq,Lal:2023dkj,Dorrill:2023vox,Forestano:2023ijh,Dersy:2023job,Cotler:2023lem,Mizera:2023bsw,Gnech:2023prs,Seong:2023njx} \end{itemize} \item \textbf{Experimental results} \\\textit{This section is incomplete as there are many results that directly and indirectly (e.g. via flavor tagging) use modern machine learning techniques. We will try to highlight experimental results that use deep learning in a critical way for the final analysis sensitivity.} diff --git a/README.md b/README.md index 84bf0b7..a61fe1f 100644 --- a/README.md +++ b/README.md @@ -74,6 +74,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [A FAIR and AI-ready Higgs Boson Decay Dataset](https://arxiv.org/abs/2108.02214) * [Particle Transformer for Jet Tagging](https://arxiv.org/abs/2202.03772) * [Public Kaggle Competition ''IceCube -- Neutrinos in Deep Ice''](https://arxiv.org/abs/2307.15289) +* [Electron Energy Regression in the CMS High-Granularity Calorimeter Prototype](https://arxiv.org/abs/2309.06582) ## Classification ### Parameterized classifiers @@ -217,6 +218,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Comparing Point Cloud Strategies for Collider Event Classification](https://arxiv.org/abs/2212.10659) * [Is infrared-collinear safe information all you need for jet classification?](https://arxiv.org/abs/2305.08979) * [Attention to Mean-Fields for Particle Cloud Generation](https://arxiv.org/abs/2305.15254) +* [A data-driven and model-agnostic approach to solving combinatorial assignment problems in searches for new physics](https://arxiv.org/abs/2309.05728) #### Physics-inspired basis @@ -309,6 +311,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Explainable Equivariant Neural Networks for Particle Physics: PELICAN](https://arxiv.org/abs/2307.16506) * [Hierarchical High-Point Energy Flow Network for Jet Tagging](https://arxiv.org/abs/2308.08300) * [Investigating the Violation of Charge-parity Symmetry Through Top-quark ChromoElectric Dipole Moments by Using Machine Learning Techniques](https://arxiv.org/abs/2306.11683) [[DOI](https://doi.org/10.5506/APhysPolB.54.5-A4)] +* [ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower \& Tracker Data Integration](https://arxiv.org/abs/2309.01568) #### strange jets @@ -630,6 +633,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Analysis of a Skyrme energy density functional with deep learning](https://arxiv.org/abs/2306.11314) * [Generative modeling of nucleon-nucleon interactions](https://arxiv.org/abs/2306.13007) * [Artificial Intelligence for the Electron Ion Collider (AI4EIC)](https://arxiv.org/abs/2307.08593) +* [Neural Network Solutions of Bosonic Quantum Systems in One Dimension](https://arxiv.org/abs/2309.02352) ### Learning strategies @@ -676,6 +680,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Symmetries, Safety, and Self-Supervision](https://arxiv.org/abs/2108.04253) * [Unsupervised Domain Transfer for Science: Exploring Deep Learning Methods for Translation between LArTPC Detector Simulations with Differing Response Models](https://arxiv.org/abs/2304.12858) * [NuCLR: Nuclear Co-Learned Representations](https://arxiv.org/abs/2306.06099) +* [A data-driven and model-agnostic approach to solving combinatorial assignment problems in searches for new physics](https://arxiv.org/abs/2309.05728) #### Reinforcement Learning @@ -982,6 +987,8 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Teaching to extract spectral densities from lattice correlators to a broad audience of learning-machines](https://arxiv.org/abs/2307.00808) * [Sampling the lattice Nambu-Goto string using Continuous Normalizing Flows](https://arxiv.org/abs/2307.01107) * [Signal-to-noise improvement through neural network contour deformations for 3D $SU(2)$ lattice gauge theory](https://arxiv.org/abs/2309.00600) +* [Application of the path optimization method to a discrete spin system](https://arxiv.org/abs/2309.06018) +* [Breaking Free with AI: The Deconfinement Transition](https://arxiv.org/abs/2309.07225) ### Function Approximation @@ -1134,6 +1141,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Toward a generative modeling analysis of CLAS exclusive $2\pi$ photoproduction](https://arxiv.org/abs/2307.04450) * [Lamarr: LHCb ultra-fast simulation based on machine learning models deployed within Gauss](https://arxiv.org/abs/2303.11428) * [SR-GAN for SR-gamma: photon super resolution at collider experiments](https://arxiv.org/abs/2308.09025) +* [CaloShowerGAN, a Generative Adversarial Networks model for fast calorimeter shower simulation](https://arxiv.org/abs/2309.06515) ### Autoencoders @@ -1213,6 +1221,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Sampling $U(1)$ gauge theory using a re-trainable conditional flow-based model](https://arxiv.org/abs/2306.00581) * [Inductive CaloFlow](https://arxiv.org/abs/2305.11934) * [SuperCalo: Calorimeter shower super-resolution](https://arxiv.org/abs/2308.11700) +* [Flows for Flows: Morphing one Dataset into another with Maximum Likelihood Estimation](https://arxiv.org/abs/2309.06472) ### Diffusion Models @@ -1230,6 +1239,8 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Renormalizing Diffusion Models](https://arxiv.org/abs/2308.12355) * [Improving Generative Model-based Unfolding with Schr\"odinger Bridges](https://arxiv.org/abs/2308.12351) * [CaloScore v2: Single-shot Calorimeter Shower Simulation with Diffusion Models](https://arxiv.org/abs/2308.03847) +* [Accelerating Markov Chain Monte Carlo sampling with diffusion models](https://arxiv.org/abs/2309.01454) +* [CaloClouds II: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation](https://arxiv.org/abs/2309.05704) ### Transformer Models @@ -1289,6 +1300,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Jet Diffusion versus JetGPT -- Modern Networks for the LHC](https://arxiv.org/abs/2305.10475) * [Implicit Quantile Networks For Emulation in Jet Physics](https://arxiv.org/abs/2306.15053) * [Towards accurate real-time luminescence thermometry: an automated machine learning approach](https://arxiv.org/abs/2307.05497) +* [ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower \& Tracker Data Integration](https://arxiv.org/abs/2309.01568) ## Anomaly detection. @@ -1475,6 +1487,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Autoencoders for Real-Time SUEP Detection](https://arxiv.org/abs/2306.13595) * [Pinning down the leptophobic $Z^\prime$ in leptonic final states with Deep Learning](https://arxiv.org/abs/2307.01118) * [Tip of the Red Giant Branch Bounds on the Neutrino Magnetic Dipole Moment Revisited](https://arxiv.org/abs/2307.13050) +* [LHC Study of Third-Generation Scalar Leptoquarks with Machine-Learned Likelihoods](https://arxiv.org/abs/2309.05407) ### Differentiable Simulation @@ -1558,6 +1571,7 @@ This review was built with the help of the HEP-ML community, the [INSPIRE REST A * [Renormalizing Diffusion Models](https://arxiv.org/abs/2308.12355) * [Scattering with Neural Operators](https://arxiv.org/abs/2308.14789) * [Distilling the essential elements of nuclear binding via neural-network quantum states](https://arxiv.org/abs/2308.16266) +* [Unsupervised Machine Learning Techniques for Exploring Tropical Coamoeba, Brane Tilings and Seiberg Duality](https://arxiv.org/abs/2309.05702) ## Experimental results. *This section is incomplete as there are many results that directly and indirectly (e.g. via flavor tagging) use modern machine learning techniques. We will try to highlight experimental results that use deep learning in a critical way for the final analysis sensitivity.* diff --git a/docs/index.md b/docs/index.md index 4ccb6f7..c78f1bf 100644 --- a/docs/index.md +++ b/docs/index.md @@ -115,6 +115,7 @@ const expandElements = shouldExpand => { * [A FAIR and AI-ready Higgs Boson Decay Dataset](https://arxiv.org/abs/2108.02214) * [Particle Transformer for Jet Tagging](https://arxiv.org/abs/2202.03772) * [Public Kaggle Competition ''IceCube -- Neutrinos in Deep Ice''](https://arxiv.org/abs/2307.15289) + * [Electron Energy Regression in the CMS High-Granularity Calorimeter Prototype](https://arxiv.org/abs/2309.06582) ## Classification @@ -268,6 +269,7 @@ const expandElements = shouldExpand => { * [Comparing Point Cloud Strategies for Collider Event Classification](https://arxiv.org/abs/2212.10659) * [Is infrared-collinear safe information all you need for jet classification?](https://arxiv.org/abs/2305.08979) * [Attention to Mean-Fields for Particle Cloud Generation](https://arxiv.org/abs/2305.15254) + * [A data-driven and model-agnostic approach to solving combinatorial assignment problems in searches for new physics](https://arxiv.org/abs/2309.05728) #### Physics-inspired basis @@ -365,6 +367,7 @@ const expandElements = shouldExpand => { * [Explainable Equivariant Neural Networks for Particle Physics: PELICAN](https://arxiv.org/abs/2307.16506) * [Hierarchical High-Point Energy Flow Network for Jet Tagging](https://arxiv.org/abs/2308.08300) * [Investigating the Violation of Charge-parity Symmetry Through Top-quark ChromoElectric Dipole Moments by Using Machine Learning Techniques](https://arxiv.org/abs/2306.11683) [[DOI](https://doi.org/10.5506/APhysPolB.54.5-A4)] + * [ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower \& Tracker Data Integration](https://arxiv.org/abs/2309.01568) #### strange jets @@ -686,6 +689,7 @@ const expandElements = shouldExpand => { * [Analysis of a Skyrme energy density functional with deep learning](https://arxiv.org/abs/2306.11314) * [Generative modeling of nucleon-nucleon interactions](https://arxiv.org/abs/2306.13007) * [Artificial Intelligence for the Electron Ion Collider (AI4EIC)](https://arxiv.org/abs/2307.08593) + * [Neural Network Solutions of Bosonic Quantum Systems in One Dimension](https://arxiv.org/abs/2309.02352) ??? example "Learning strategies" @@ -737,6 +741,7 @@ const expandElements = shouldExpand => { * [Symmetries, Safety, and Self-Supervision](https://arxiv.org/abs/2108.04253) * [Unsupervised Domain Transfer for Science: Exploring Deep Learning Methods for Translation between LArTPC Detector Simulations with Differing Response Models](https://arxiv.org/abs/2304.12858) * [NuCLR: Nuclear Co-Learned Representations](https://arxiv.org/abs/2306.06099) + * [A data-driven and model-agnostic approach to solving combinatorial assignment problems in searches for new physics](https://arxiv.org/abs/2309.05728) #### Reinforcement Learning @@ -1083,6 +1088,8 @@ const expandElements = shouldExpand => { * [Teaching to extract spectral densities from lattice correlators to a broad audience of learning-machines](https://arxiv.org/abs/2307.00808) * [Sampling the lattice Nambu-Goto string using Continuous Normalizing Flows](https://arxiv.org/abs/2307.01107) * [Signal-to-noise improvement through neural network contour deformations for 3D $SU(2)$ lattice gauge theory](https://arxiv.org/abs/2309.00600) + * [Application of the path optimization method to a discrete spin system](https://arxiv.org/abs/2309.06018) + * [Breaking Free with AI: The Deconfinement Transition](https://arxiv.org/abs/2309.07225) ??? example "Function Approximation" @@ -1261,6 +1268,7 @@ const expandElements = shouldExpand => { * [Toward a generative modeling analysis of CLAS exclusive $2\pi$ photoproduction](https://arxiv.org/abs/2307.04450) * [Lamarr: LHCb ultra-fast simulation based on machine learning models deployed within Gauss](https://arxiv.org/abs/2303.11428) * [SR-GAN for SR-gamma: photon super resolution at collider experiments](https://arxiv.org/abs/2308.09025) + * [CaloShowerGAN, a Generative Adversarial Networks model for fast calorimeter shower simulation](https://arxiv.org/abs/2309.06515) ??? example "Autoencoders" @@ -1350,6 +1358,7 @@ const expandElements = shouldExpand => { * [Sampling $U(1)$ gauge theory using a re-trainable conditional flow-based model](https://arxiv.org/abs/2306.00581) * [Inductive CaloFlow](https://arxiv.org/abs/2305.11934) * [SuperCalo: Calorimeter shower super-resolution](https://arxiv.org/abs/2308.11700) + * [Flows for Flows: Morphing one Dataset into another with Maximum Likelihood Estimation](https://arxiv.org/abs/2309.06472) ??? example "Diffusion Models" @@ -1372,6 +1381,8 @@ const expandElements = shouldExpand => { * [Renormalizing Diffusion Models](https://arxiv.org/abs/2308.12355) * [Improving Generative Model-based Unfolding with Schr\"odinger Bridges](https://arxiv.org/abs/2308.12351) * [CaloScore v2: Single-shot Calorimeter Shower Simulation with Diffusion Models](https://arxiv.org/abs/2308.03847) + * [Accelerating Markov Chain Monte Carlo sampling with diffusion models](https://arxiv.org/abs/2309.01454) + * [CaloClouds II: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation](https://arxiv.org/abs/2309.05704) ??? example "Transformer Models" @@ -1461,6 +1472,7 @@ const expandElements = shouldExpand => { * [Jet Diffusion versus JetGPT -- Modern Networks for the LHC](https://arxiv.org/abs/2305.10475) * [Implicit Quantile Networks For Emulation in Jet Physics](https://arxiv.org/abs/2306.15053) * [Towards accurate real-time luminescence thermometry: an automated machine learning approach](https://arxiv.org/abs/2307.05497) + * [ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower \& Tracker Data Integration](https://arxiv.org/abs/2309.01568) ## Anomaly detection. @@ -1670,6 +1682,7 @@ const expandElements = shouldExpand => { * [Autoencoders for Real-Time SUEP Detection](https://arxiv.org/abs/2306.13595) * [Pinning down the leptophobic $Z^\prime$ in leptonic final states with Deep Learning](https://arxiv.org/abs/2307.01118) * [Tip of the Red Giant Branch Bounds on the Neutrino Magnetic Dipole Moment Revisited](https://arxiv.org/abs/2307.13050) + * [LHC Study of Third-Generation Scalar Leptoquarks with Machine-Learned Likelihoods](https://arxiv.org/abs/2309.05407) ??? example "Differentiable Simulation" @@ -1788,6 +1801,7 @@ const expandElements = shouldExpand => { * [Renormalizing Diffusion Models](https://arxiv.org/abs/2308.12355) * [Scattering with Neural Operators](https://arxiv.org/abs/2308.14789) * [Distilling the essential elements of nuclear binding via neural-network quantum states](https://arxiv.org/abs/2308.16266) + * [Unsupervised Machine Learning Techniques for Exploring Tropical Coamoeba, Brane Tilings and Seiberg Duality](https://arxiv.org/abs/2309.05702) ## Experimental results. *This section is incomplete as there are many results that directly and indirectly (e.g. via flavor tagging) use modern machine learning techniques. We will try to highlight experimental results that use deep learning in a critical way for the final analysis sensitivity.* diff --git a/docs/recent.md b/docs/recent.md index 5c36b26..b7b12e7 100644 --- a/docs/recent.md +++ b/docs/recent.md @@ -10,6 +10,18 @@ search: This is an automatically compiled list of papers which have been added to the living review that were made public within the previous 4 months at the time of updating. This is not an exhaustive list of released papers, and is only able to find those which have both year and month data provided in the bib reference. ## September 2023 +* [Breaking Free with AI: The Deconfinement Transition](https://arxiv.org/abs/2309.07225) +* [CaloShowerGAN, a Generative Adversarial Networks model for fast calorimeter shower simulation](https://arxiv.org/abs/2309.06515) +* [Electron Energy Regression in the CMS High-Granularity Calorimeter Prototype](https://arxiv.org/abs/2309.06582) +* [Flows for Flows: Morphing one Dataset into another with Maximum Likelihood Estimation](https://arxiv.org/abs/2309.06472) +* [CaloClouds II: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation](https://arxiv.org/abs/2309.05704) +* [Unsupervised Machine Learning Techniques for Exploring Tropical Coamoeba, Brane Tilings and Seiberg Duality](https://arxiv.org/abs/2309.05702) +* [Application of the path optimization method to a discrete spin system](https://arxiv.org/abs/2309.06018) +* [A data-driven and model-agnostic approach to solving combinatorial assignment problems in searches for new physics](https://arxiv.org/abs/2309.05728) +* [LHC Study of Third-Generation Scalar Leptoquarks with Machine-Learned Likelihoods](https://arxiv.org/abs/2309.05407) +* [Neural Network Solutions of Bosonic Quantum Systems in One Dimension](https://arxiv.org/abs/2309.02352) +* [Accelerating Markov Chain Monte Carlo sampling with diffusion models](https://arxiv.org/abs/2309.01454) +* [ML-Based Top Taggers: Performance, Uncertainty and Impact of Tower \& Tracker Data Integration](https://arxiv.org/abs/2309.01568) * [Signal-to-noise improvement through neural network contour deformations for 3D $SU(2)$ lattice gauge theory](https://arxiv.org/abs/2309.00600) ## August 2023