google,我,萨娜

By 方华. 2006

9,238 citations recorded by the source corpus.

Subfields: Natural Language Processing, Computational Linguistics & Semantics

How researchers see this paper

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

In Natural Language Processing

In Computational Linguistics & Semantics

Citations in the corpus

Pathways: Asynchronous Distributed Dataflow for ML (2022, neutral): “Gang-scheduling is essential in the case of TPUs, since they are single-threaded and only run non-preemptible kernels, so the system will deadlock if communicating computations are not enqueued in a consistent order.”
Efficiently Scaling Transformer Inference (2022, neutral): “We use up to 256 TPU v4 chips (Google, 2022) for our benchmarks.”
Extracting Training Data from Large Language Models (2020, background): “Private machine learning models can be trained with variants of the differentially private stochastic gradient descent (DP-SGD) algorithm [1] which is widely implemented [17,25].”
Quantifying the Carbon Emissions of Machine Learning (2019, background): “For instance, Google Cloud Platform is certiﬁed carbon neutral and funds solar and wind farms directly on local grids through RECs [13].”
Fact, Fetch, and Reason: A Unified Evaluation of Retrieval-Augmented Generation (2024, background): “This piecemeal approach fails to capture the holistic performance of these systems in real-world applications (Yu et al., 2024b).”
MaskGIT: Masked Generative Image Transformer (2022, methodology): “…editing tasks: class-conditional evaluated on the corresponding 256 ˆ 256 ones, consistent with their training; ˝ taken from the prior work; : evaluated using the released model trained on a subset of Places2; ˛ evaluated using the TFHub model [18]. image editing, image inpainting, and outpainting.”
Certified Robustness to Adversarial Examples with Differential Privacy (2018, methodology): “ImageNet: We use as baseline a pre-trained version of Inception-v3 [53] available in Tensorﬂow [20].”
DRN: A Deep Reinforcement Learning Framework for News Recommendation (2018, background): “For instance, one of the most popular on-line services, news aggregation services, such as Google News [15] can provide overwhelming volume of content than the amount that users can digest.”
Practical and Private (Deep) Learning without Sampling or Shuffling (2021, methodology): “Similarly, for DP-SGD [31], we consider its momentum variant (DP-SGDM), and report the best-performing variant in each task.”
Secure Evaluation of Quantized Neural Networks (2019, methodology): “…on a different scheme to get efﬁcient inference) we choose this particular scheme for the following reasons: It is implemented in Tensorﬂow (more precisely, TFLite [28]) and as such we get an user friendly, widely available and well documented way of training models that can securely evaluated.”
FairSight: Visual Analytics for Fairness in Decision Making (2019, methodology): “With the increasing awareness of the importance of fair machine learning, a number of new tools, including API [3] and interface [40], as well as integrated systems such as the What-if tool [15] or AI Fairness 360 [22], have been developed.”
Neural Combinatorial Optimization with Reinforcement Learning (2016, methodology): “In contrast, machine learning methods have the potential to be applicable across many optimization tasks by automatically discovering their own heuristics based on the training data, thus requiring less hand-engineering than solvers that are optimized for one task only.”
Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference (2017, background): “One can see in TensorFlow Lite [5] how these are called.”
Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need? (2018, background): “The budding area of “UX for ML” has begun to explore new forms of prototyping for ML systems, to provide earlier insights into the complex, interacting UX impacts of particular data, modeling, and system-design choices [27, 41, 47, 97].”
Gmail Smart Compose: Real-Time Assisted Writing (2019, methodology): “By leveraging the latest deep learning model inference technologies [17, 25], our Smart Compose system gives one example of how neural language models can be served at industry-scale for billions of users.”