Recent success in deep learning has partially been driven by training increasingly
overparametrized networks on ever larger datasets. It is therefore natural to ask: how much …

Instruction tuning plays a critical role in aligning large language models (LLMs) with human
preference. Despite the vast amount of open instruction datasets, naively training a LLM on …

Machine learning training data is often dynamic in real-world use cases, ie, data is added or
removed and may experience distribution shifts over time. Models must incorporate this …

At its core, this thesis aims to enhance the practicality of deep learning by improving the
label and training efficiency of deep learning models. To this end, we investigate data subset …

Quantization-aware training (QAT) is a representative model compression method to reduce
redundancy in weights and activations. However, most existing QAT methods require end-to …

In real-world machine learning (ML) pipelines, datasets are continuously growing. Models
must incorporate this new training data to improve generalization and adapt to potential …

Quantization-aware training (QAT) is a representative model compression method to reduce
redundancy in weights and activations. However, most existing QAT methods require end-to …