Zobrazeno 1 - 10
of 86
pro vyhledávání: '"Dubois, Yann"'
Autor:
Sun, Yu, Li, Xinhao, Dalal, Karan, Xu, Jiarui, Vikram, Arjun, Zhang, Genghan, Dubois, Yann, Chen, Xinlei, Wang, Xiaolong, Koyejo, Sanmi, Hashimoto, Tatsunori, Guestrin, Carlos
Self-attention performs well in long context but has quadratic complexity. Existing RNN layers have linear complexity, but their performance in long context is limited by the expressive power of their hidden state. We propose a new class of sequence
Externí odkaz:
http://arxiv.org/abs/2407.04620
LLM-based auto-annotators have become a key component of the LLM development process due to their cost-effectiveness and scalability compared to human-based evaluation. However, these auto-annotators can introduce complex biases that are hard to remo
Externí odkaz:
http://arxiv.org/abs/2404.04475
Autor:
Ruan, Yangjun, Dong, Honghua, Wang, Andrew, Pitis, Silviu, Zhou, Yongchao, Ba, Jimmy, Dubois, Yann, Maddison, Chris J., Hashimoto, Tatsunori
Recent advances in Language Model (LM) agents and tool use, exemplified by applications like ChatGPT Plugins, enable a rich set of capabilities but also amplify potential risks - such as leaking private data or causing financial losses. Identifying t
Externí odkaz:
http://arxiv.org/abs/2309.15817
Autor:
Dubois, Yann, Li, Xuechen, Taori, Rohan, Zhang, Tianyi, Gulrajani, Ishaan, Ba, Jimmy, Guestrin, Carlos, Liang, Percy, Hashimoto, Tatsunori B.
Large language models (LLMs) such as ChatGPT have seen widespread adoption due to their strong instruction-following abilities. Developing these LLMs involves a complex yet poorly understood workflow requiring training with human feedback. Replicatin
Externí odkaz:
http://arxiv.org/abs/2305.14387
Self-supervised learning (SSL) pipelines differ in many design choices such as the architecture, augmentations, or pretraining data. Yet SSL is typically evaluated using a single metric: linear probing on ImageNet. This does not provide much insight
Externí odkaz:
http://arxiv.org/abs/2302.03068
Despite the empirical successes of self-supervised learning (SSL) methods, it is unclear what characteristics of their representations lead to high downstream accuracies. In this work, we characterize properties that SSL representations should ideall
Externí odkaz:
http://arxiv.org/abs/2209.06235
The development of CLIP [Radford et al., 2021] has sparked a debate on whether language supervision can result in vision models with more transferable representations than traditional image-only methods. Our work studies this question through a caref
Externí odkaz:
http://arxiv.org/abs/2207.07635
Autor:
Miao, Ning, Rainforth, Tom, Mathieu, Emile, Dubois, Yann, Teh, Yee Whye, Foster, Adam, Kim, Hyunjik
We introduce InstaAug, a method for automatically learning input-specific augmentations from data. Previous methods for learning augmentations have typically assumed independence between the original input and the transformation applied to that input
Externí odkaz:
http://arxiv.org/abs/2206.00051
Machine learning systems often experience a distribution shift between training and testing. In this paper, we introduce a simple variational objective whose optima are exactly the set of all representations on which risk minimizers are guaranteed to
Externí odkaz:
http://arxiv.org/abs/2201.00057
Most data is automatically collected and only ever "seen" by algorithms. Yet, data compressors preserve perceptual fidelity rather than just the information needed by algorithms performing downstream tasks. In this paper, we characterize the bit-rate
Externí odkaz:
http://arxiv.org/abs/2106.10800