Reinforcement learning from human feedback (RLHF) utilizes human feedback to better align large language models with human preferences via online optimization against a learned reward model. Current RLHF paradigms rely on Proximal Policy Optimization (PPO), which quickly becomes a challenge to implement and scale up to large architectures. To address this difficulty we present the AutoRLHF library as a feature complete open-source framework for RLHF fine-tuning of models up to and exceeding 70 billion parameters. To do so we implement support for multiple types of distributed training including distributed data parallel, model sharded, as well as tensor, sequential, and pipeline parallelism. Additionally, we implement compute and memory saving features, giving AutoRLHF the flexibility to support users with a wide range of compute resources. This includes offline RL methods like Implicit Language Q Learning (ILQL) as a compute efficient alternative to PPO. We find offline fine-tuning offers competitive performance relative to online algorithms while being easier to implement, train, and scale. To evaluate our framework we train RLHF models on two separate well-known tasks using publicly available human preference data. Models trained with AutoRLHF achieve preference win-rates over baselines at rates comparable to the original works.

Low-quality data can cause downstream problems in high-stakes applications. Data-centric approach emphasizes on improving dataset quality to enhance model performance. High-quality datasets are needed for general-purpose Large Language Models (LLMs) training, as well as for domain-specific models, which are usually small in size as it is costly to engage a large number of domain experts for their creation. Thus, it is vital to ensure high-quality domain-specific training data. In this paper, we propose a framework for enhancing the data quality of original datasets. (Code and dataset are available at https://github.com/IvaBojic/framework). We applied the proposed framework to four biomedical datasets and showed relative improvement of up to 33%/40% for fine-tuning of retrieval/reader models on the BioASQ dataset when using back translation to enhance the original dataset quality.

We study a new problem of cross-lingual transfer learning for event coreference resolution (ECR) where models trained on data from a source language are adapted for evaluations in different target languages. We introduce the first baseline model for this task based on XLM-RoBERTa, a state-of-the-art multilingual pre-trained language model. We also explore language adversarial neural networks (LANN) that present language discriminators to distinguish texts from the source and target languages to improve the language generalization for ECR. In addition, we introduce two novel mechanisms to further enhance the general representation learning of LANN, featuring: (i) multi-view alignment to penalize cross coreference-label alignment of examples in the source and target languages, and (ii) optimal transport to select close examples in the source and target languages to provide better training signals for the language discriminators. Finally, we perform extensive experiments for cross-lingual ECR from English to Spanish and Chinese to demonstrate the effectiveness of the proposed methods.

We study the problem of event coreference resolution (ECR) that seeks to group coreferent event mentions into the same clusters. Deep learning methods have recently been applied for this task to deliver state-of-the-art performance. However, existing deep learning models for ECR are limited in that they cannot exploit important interactions between relevant objects for ECR, e.g., context words and entity mentions, to support the encoding of document-level context. In addition, consistency constraints between golden and predicted clusters of event mentions have not been considered to improve representation learning in prior deep learning models for ECR. This work addresses such limitations by introducing a novel deep learning model for ECR. At the core of our model are document structures to explicitly capture relevant objects for ECR. Our document structures introduce diverse knowledge sources (discourse, syntax, semantics) to compute edges/interactions between structure nodes for document-level representation learning. We also present novel regularization techniques based on consistencies of golden and predicted clusters for event mentions in documents. Extensive experiments show that our model achieve state-of-the-art performance on two benchmark datasets.

This paper studies the problem of cross-document event coreference resolution (CDECR) that seeks to determine if event mentions across multiple documents refer to the same real-world events. Prior work has demonstrated the benefits of the predicate-argument information and document context for resolving the coreference of event mentions. However, such information has not been captured effectively in prior work for CDECR. To address these limitations, we propose a novel deep learning model for CDECR that introduces hierarchical graph convolutional neural networks (GCN) to jointly resolve entity and event mentions. As such, sentence-level GCNs enable the encoding of important context words for event mentions and their arguments while the document-level GCN leverages the interaction structures of event mentions and arguments to compute document representations to perform CDECR. Extensive experiments are conducted to demonstrate the effectiveness of the proposed model.

Personality image captioning (PIC) aims to describe an image with a natural language caption given a personality trait. In this work, we introduce a novel formulation for PIC based on a communication game between a speaker and a listener. The speaker attempts to generate natural language captions while the listener encourages the generated captions to contain discriminative information about the input images and personality traits. In this way, we expect that the generated captions can be improved to naturally represent the images and express the traits. In addition, we propose to adapt the language model GPT2 to perform caption generation for PIC. This enables the speaker and listener to benefit from the language encoding capacity of GPT2. Our experiments show that the proposed model achieves the state-of-the-art performance for PIC.