A and V

By R. Stephenson. 1962

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Multimodal Learning

Citations in the corpus

Very Deep Convolutional Networks for Large-Scale Image Recognition (2014, background): “visual recognition ConvNet, applied to large datasets, is t hat of [5], who addressed the task of street number recognition.”
The Capacity for Moral Self-Correction in Large Language Models (2023, background): “They find that the model they test—RoBERTa-large (345M parameters) [34]3—does not produce less biased outputs when instructed to do so with natural language interventions.”
Designing Network Design Spaces (2020, methodology): “di = [1, 1, 4, 7] wi = [24, 56, 152, 368] g = 8, b = 1, e = 30.”
Quantized Convolutional Neural Networks for Mobile Devices (2015, methodology): “From Table 4, we discover that with a large speed-up rate (over 4 × ), the performance loss of both CPD and GBD become severe, especially before ﬁne-tuning.”
Byzantine-Resilient Secure Federated Learning (2020, background): “Another important direction in federated learning is the study of fairness and how to avoid biasing the model towards specific users [28], [29].”
Accelerating Very Deep Convolutional Networks for Classification and Detection (2015, methodology): “We observe that the learning 582 rate needs to be carefully chosen for the SGD-based 583 data reconstruction to converge (as also reported 584 independently in [18] for another decomposition), 585 and when the training starts to converge, the results 586 are still sensitive to the initialization (fo”
Gaussian approximation potentials: the accuracy of quantum mechanics, without the electrons. (2009, background): “The advantage of this is that the 4D surface contains all the information from the 3D spherical region inside the cutoff, including the radial dimension, and thus 4D spherical harmonics (also called Wigner matrices, U j m′m[11]) constitute a natural complete basis for the interior of the 3D sphere, ”
Wasserstein Auto-Encoders (2017, background): “The approach proposed by [19] is perhaps the most relevant to our work.”
Cyberbullying Detection With Fairness Constraints (2020, background): “detect bias-sensitive words and replace them with neutral words or tokens to reduce stereotypical bias for hate speech detection [49].”
Regularity Criteria for the Three-dimensional Navier-Stokes Equations (2008, background): “Most recently, there has been some progress along these lines (see, for example, [2], [3], [15], [18], [19], [21], [41], [42], and references therein) which states, roughly speaking, that a strong solution u exists on the time interval [0, T] for as long as”
DFA on AES (2004, background): “For more information, the reader can refer to [16, 10].”
Content-Based Music Information Retrieval: Current Directions and Future Challenges (2008, methodology): “One approach to reducing the time complexity of pairwise distance computation that has much potential is locality sensitive hashing (LSH), which trades time complexity of the query for space complexity of the database and a possible slight loss of accuracy [118], [119].”
Worldwide Increasing Incidence of Thyroid Cancer: Update on Epidemiology and Risk Factors (2013, methodology): “A recentmeta-analysis supported the inference that thyroid cancer is less frequent in MNG than in SN, although this nding appears to hold true mostly outside the United States, in iodine-de cient populations [70].”
Heteroatom-doped graphene materials: syntheses, properties and applications. (2014, background): “9b).(117) It has been predicted that the Fermi level decreases ~0.”
Ultracold molecules and ultracold chemistry (2009, background): “The observation of a charge transfer reaction between laser-cooled Ybþ ions in a Paul trap and Yb atoms in a magneto-optical trap has recently been reported [237].”
Review: Hydrothermal technology for smart materials (2016, background): “Materials such as PbTiO3 and hydroxyapatite have been made with this approach.(35,36) It is known that high pressures in excess of 1 GPa catalyse low-temperature solid-state reactions in ceramic materials by orders of magnitude.”