相关论文: How to Leverage Demonstration Data in Alignment fo…

In-Context Learning (ICL) combined with pre-trained large language models has achieved promising results on various NLP tasks. However, ICL requires high-quality annotated demonstrations which might not be available in real-world scenarios.…

State-of-the-art frameworks in self-supervised learning have recently shown that fully utilizing transformer-based models can lead to performance boost compared to conventional CNN models. Striving to maximize the mutual information of two…

All text-based language problems can be reduced to either generation or embedding. Current models only perform well at one or the other. We introduce generative representational instruction tuning (GRIT) whereby a large language model is…

Fine-tuning large language models (LLMs) for alignment typically relies on supervised fine-tuning or reinforcement learning from human feedback, both limited by the cost and scarcity of high-quality annotations. Recent self-play and…

Language models are aligned to emulate the collective voice of many, resulting in outputs that align with no one in particular. Steering LLMs away from generic output is possible through supervised finetuning or RLHF, but requires…

How well are unimodal vision and language models aligned? Although prior work have approached answering this question, their assessment methods do not directly translate to how these models are used in practical vision-language tasks. In…

Meta-learning that uses implicit gradient have provided an exciting alternative to standard techniques which depend on the trajectory of the inner loop training. Implicit meta-learning (IML), however, require computing $2^{nd}$ order…

We study the problem of sample efficient reinforcement learning, where prior data such as demonstrations are provided for initialization in lieu of a dense reward signal. A natural approach is to incorporate an imitation learning objective,…

Self-supervised Learning (SSL) including the mainstream contrastive learning has achieved great success in learning visual representations without data annotations. However, most of methods mainly focus on the instance level information…

Large Language Models (LLMs) have achieved remarkable success across a wide range of natural language tasks, and recent efforts have sought to extend their capabilities to multimodal domains and resource-constrained environments. However,…

The introduction of the generative adversarial imitation learning (GAIL) algorithm has spurred the development of scalable imitation learning approaches using deep neural networks. Many of the algorithms that followed used a similar…

Systematic Generalization refers to a learning algorithm's ability to extrapolate learned behavior to unseen situations that are distinct but semantically similar to its training data. As shown in recent work, state-of-the-art deep learning…

Today's autonomous agents, largely driven by foundation models (FMs), can understand natural language instructions and solve long-horizon tasks with human-like reasoning. However, current human-robot interaction largely follows a one-way…

In-context learning (ICL) enables large language models (LLMs) to adapt to new tasks during inference using only a few demonstrations. However, ICL performance is highly dependent on the selection of these demonstrations. Recent work…

Adversarial Imitation Learning (AIL) is a broad family of imitation learning methods designed to mimic expert behaviors from demonstrations. While AIL has shown state-of-the-art performance on imitation learning with only small number of…

In robotic manipulation, acquiring samples is extremely expensive because it often requires interacting with the real world. Traditional image-level data augmentation has shown the potential to improve sample efficiency in various machine…

Despite the numerous breakthroughs achieved with Reinforcement Learning (RL), solving environments with sparse rewards remains a challenging task that requires sophisticated exploration. Learning from Demonstrations (LfD) remedies this…

Traditional implicit generative models are capable of learning highly complex data distributions. However, their training involves distinguishing real data from synthetically generated data using adversarial discriminators, which can lead…

Large language models (LLMs) exhibit remarkable capabilities across diverse tasks, yet aligning them efficiently and effectively with human expectations remains a critical challenge. This thesis advances LLM alignment by introducing novel…

In-context learning (ICL), teaching a large language model (LLM) to perform a task with few-shot demonstrations rather than adjusting the model parameters, has emerged as a strong paradigm for using LLMs. While early studies primarily used…