TrAct: Making First-layer Pre-Activations Trainable

Autor: Petersen, Felix, Borgelt, Christian, Ermon, Stefano

We consider the training of the first layer of vision models and notice the clear relationship between pixel values and gradient update magnitudes: the gradients arriving at the weights of a first layer are by definition directly proportional to (nor

Externí odkaz: http://arxiv.org/abs/2410.23970

Newton Losses: Using Curvature Information for Learning with Differentiable Algorithms

Autor: Petersen, Felix, Borgelt, Christian, Sutter, Tobias, Kuehne, Hilde, Deussen, Oliver, Ermon, Stefano

When training neural networks with custom objectives, such as ranking losses and shortest-path losses, a common problem is that they are, per se, non-differentiable. A popular approach is to continuously relax the objectives to provide gradients, ena

Externí odkaz: http://arxiv.org/abs/2410.19055

Electric Grid Topology and Admittance Estimation: Quantifying Phasor-based Measurement Requirements

Autor: Rin, Norak, Shames, Iman, Petersen, Ian R., Ratnam, Elizabeth L.

In this paper, we quantify voltage and current phasor-based measurement requirements for the unique identification of the electric grid topology and admittance parameters. Our approach is underpinned by the concept of a rigidity matrix that has been

Externí odkaz: http://arxiv.org/abs/2410.17553

Shining Light on the Dark Sector: Search for Axion-like Particles and Other New Physics in Photonic Final States with FASER

Autor: FASER collaboration, Abraham, Roshan Mammen, Ai, Xiaocong, Anders, John, Antel, Claire, Ariga, Akitaka, Ariga, Tomoko, Atkinson, Jeremy, Bernlochner, Florian U., Bianchi, Emma, Boeckh, Tobias, Boyd, Jamie, Brenner, Lydia, Burger, Angela, Cadoux, Franck, Cardella, Roberto, Casper, David W., Cavanagh, Charlotte, Chen, Xin, Cho, Eunhyung, Chouhan, Dhruv, Coccaro, Andrea, Débieux, Stephane, D'Onofrio, Monica, Desai, Ansh, Dmitrievsky, Sergey, Dobre, Radu, Eley, Sinead, Favre, Yannick, Fellers, Deion, Feng, Jonathan L., Fenoglio, Carlo Alberto, Ferrere, Didier, Fieg, Max, Filali, Wissal, Firu, Elena, Garabaglu, Ali, Gibson, Stephen, Gonzalez-Sevilla, Sergio, Gornushkin, Yuri, Gwilliam, Carl, Hayakawa, Daiki, Holzbock, Michael, Hsu, Shih-Chieh, Hu, Zhen, Iacobucci, Giuseppe, Inada, Tomohiro, Iodice, Luca, Jakobsen, Sune, Joos, Hans, Kajomovitz, Enrique, Kawahara, Hiroaki, Keyken, Alex, Kling, Felix, Köck, Daniela, Kontaxakis, Pantelis, Kose, Umut, Kotitsa, Rafaella, Kuehn, Susanne, Kugathasan, Thanushan, Levinson, Lorne, Li, Ke, Liu, Jinfeng, Liu, Yi, Lutz, Margaret S., MacDonald, Jack, Magliocca, Chiara, Mäkelä, Toni, McCoy, Lawson, McFayden, Josh, Medina, Andrea Pizarro, Milanesio, Matteo, Moretti, Théo, Nakamura, Mitsuhiro, Nakano, Toshiyuki, Nevay, Laurie, Ohashi, Ken, Otono, Hidetoshi, Paolozzi, Lorenzo, Petersen, Brian, Preda, Titi, Prim, Markus, Queitsch-Maitland, Michaela, Rokujo, Hiroki, Rubbia, André, Sabater-Iglesias, Jorge, Sato, Osamu, Scampoli, Paola, Schmieden, Kristof, Schott, Matthias, Sfyrla, Anna, Sgalaberna, Davide, Shamim, Mansoora, Shively, Savannah, Takubo, Yosuke, Tarannum, Noshin, Theiner, Ondrej, Torrence, Eric, Martinez, Oscar Ivan Valdes, Vasina, Svetlana, Vormwald, Benedikt, Wang, Di, Wang, Yuxiao, Welch, Eli, Xu, Yue, Zahorec, Samuel, Zambito, Stefano, Zhang, Shunliang

The first FASER search for a light, long-lived particle decaying into a pair of photons is reported. The search uses LHC proton-proton collision data at $\sqrt{s}=13.6~\text{TeV}$ collected in 2022 and 2023, corresponding to an integrated luminosity

Externí odkaz: http://arxiv.org/abs/2410.10363

Generalizing Stochastic Smoothing for Differentiation and Gradient Estimation

Autor: Petersen, Felix, Borgelt, Christian, Mishra, Aashwin, Ermon, Stefano

We deal with the problem of gradient estimation for stochastic differentiable relaxations of algorithms, operators, simulators, and other non-differentiable functions. Stochastic smoothing conventionally perturbs the input of a non-differentiable fun

Externí odkaz: http://arxiv.org/abs/2410.08125