Zobrazeno 1 - 10
of 1 211
pro vyhledávání: '"MAHONEY, MICHAEL"'
Autor:
Masserano, Luca, Ansari, Abdul Fatir, Han, Boran, Zhang, Xiyuan, Faloutsos, Christos, Mahoney, Michael W., Wilson, Andrew Gordon, Park, Youngsuk, Rangapuram, Syama, Maddix, Danielle C., Wang, Yuyang
How to best develop foundational models for time series forecasting remains an important open question. Tokenization is a crucial consideration in this effort: what is an effective discrete vocabulary for a real-valued sequential input? To address th
Externí odkaz:
http://arxiv.org/abs/2412.05244
Autor:
Zhang, Hanyu, Arvin, Chuck, Efimov, Dmitry, Mahoney, Michael W., Perrault-Joncas, Dominique, Ramasubramanian, Shankar, Wilson, Andrew Gordon, Wolff, Malcolm
Modern time-series forecasting models often fail to make full use of rich unstructured information about the time series themselves. This lack of proper conditioning can lead to obvious model failures; for example, models may be unaware of the detail
Externí odkaz:
http://arxiv.org/abs/2412.02525
Autor:
Cheng, Chaoran, Han, Boran, Maddix, Danielle C., Ansari, Abdul Fatir, Stuart, Andrew, Mahoney, Michael W., Wang, Yuyang
Generative models that satisfy hard constraints are crucial in many scientific and engineering applications where physical laws or system requirements must be strictly respected. However, many existing constrained generative models, especially those
Externí odkaz:
http://arxiv.org/abs/2412.01786
Autor:
Geniesse, Caleb, Chen, Jiaqing, Xie, Tiankai, Shi, Ge, Yang, Yaoqing, Morozov, Dmitriy, Perciano, Talita, Mahoney, Michael W., Maciejewski, Ross, Weber, Gunther H.
In machine learning, a loss function measures the difference between model predictions and ground-truth (or target) values. For neural network models, visualizing how this loss changes as model parameters are varied can provide insights into the loca
Externí odkaz:
http://arxiv.org/abs/2411.12136
Autor:
Xie, Tiankai, Geniesse, Caleb, Chen, Jiaqing, Yang, Yaoqing, Morozov, Dmitriy, Mahoney, Michael W., Maciejewski, Ross, Weber, Gunther H.
Characterizing the loss of a neural network with respect to model parameters, i.e., the loss landscape, can provide valuable insights into properties of that model. Various methods for visualizing loss landscapes have been proposed, but less emphasis
Externí odkaz:
http://arxiv.org/abs/2411.09807
Autor:
Hooper, Coleman, Kim, Sehoon, Mohammadzadeh, Hiva, Maheswaran, Monishwaran, Paik, June, Mahoney, Michael W., Keutzer, Kurt, Gholami, Amir
Emerging Large Language Model (LLM) applications require long input prompts to perform complex downstream tasks like document analysis and code generation. For these long context length applications, the length of the input prompt poses a significant
Externí odkaz:
http://arxiv.org/abs/2411.09688
Autor:
Wolff, Malcolm, Olivares, Kin G., Oreshkin, Boris, Ruan, Sunny, Yang, Sitan, Katoch, Abhinav, Ramasubramanian, Shankar, Zhang, Youxin, Mahoney, Michael W., Efimov, Dmitry, Quenneville-Bélair, Vincent
Publikováno v:
In 31st Conference on Neural Information Processing In 38th Conference on Neural Information Processing Systems NIPS 2017, Time Series in the Age of Large Models Workshop, 2024
Demand forecasting faces challenges induced by Peak Events (PEs) corresponding to special periods such as promotions and holidays. Peak events create significant spikes in demand followed by demand ramp down periods. Neural networks like MQCNN and MQ
Externí odkaz:
http://arxiv.org/abs/2411.05852
As performance gains through scaling data and/or model size experience diminishing returns, it is becoming increasingly popular to turn to ensembling, where the predictions of multiple models are combined to improve accuracy. In this paper, we provid
Externí odkaz:
http://arxiv.org/abs/2411.00328
Recent work on pruning large language models (LLMs) has shown that one can eliminate a large number of parameters without compromising performance, making pruning a promising strategy to reduce LLM model size. Existing LLM pruning strategies typicall
Externí odkaz:
http://arxiv.org/abs/2410.10912
Autor:
Lim, Soon Hoe, Wang, Yijin, Yu, Annan, Hart, Emma, Mahoney, Michael W., Li, Xiaoye S., Erichson, N. Benjamin
Flow matching has recently emerged as a powerful paradigm for generative modeling and has been extended to probabilistic time series forecasting in latent spaces. However, the impact of the specific choice of probability path model on forecasting per
Externí odkaz:
http://arxiv.org/abs/2410.03229