Fenia Christopoulou

Preference Optimization (PO) has proven an effective step for aligning language models to human-desired behaviors. Current variants, following the offline Direct Preference Optimization objective, have focused on a strict setting where all tokens are contributing signals of KL divergence and rewards to the loss function. However, human preference is not affected by each word in a sequence equally but is often dependent on specific words or phrases, e.g. existence of toxic terms leads to non-preferred responses. Based on this observation, we argue that not all tokens should be weighted equally during PO and propose a flexible objective termed SparsePO, that aims to automatically learn to weight the KL divergence and reward corresponding to each token during PO training. We propose two different variants of weight-masks that can either be derived from the reference model itself or learned on the fly. Notably, our method induces sparsity in the learned masks, allowing the model to learn how to best weight reward and KL divergence contributions at the token level, learning an optimal level of mask sparsity. Extensive experiments on multiple domains, including sentiment control, dialogue, text summarization and text-to-code generation, illustrate that our approach assigns meaningful weights to tokens according to the target task, generates more responses with the desired preference and improves reasoning tasks by up to 2 percentage points compared to other token- and response-level PO methods.

Large pre-trained language models have recently been expanded and applied to programming language tasks with great success, often through further pre-training of a strictly-natural language model–where training sequences typically contain both natural and (linearised) programming language. Such approaches effectively map both modalities of the sequence into the same embedding space. However, programming language keywords (e.g. “while”) often have very strictly defined semantics. As such, transfer learning from their natural language usage may not necessarily be beneficial to their code application and vise versa. Assuming an already pre-trained language model, in this work we investigate how sequence tokens can be adapted and represented differently, depending on which modality they belong to, and to the ultimate benefit of the downstream task. We experiment with separating embedding spaces between modalities during further model pre-training with modality-relative training objectives. We focus on text-to-code generation and observe consistent improvements across two backbone models and two test sets, measuring pass@k and a novel incremental variation.

We propose a multi-task, probabilistic approach to facilitate distantly supervised relation extraction by bringing closer the representations of sentences that contain the same Knowledge Base pairs. To achieve this, we bias the latent space of sentences via a Variational Autoencoder (VAE) that is trained jointly with a relation classifier. The latent code guides the pair representations and influences sentence reconstruction. Experimental results on two datasets created via distant supervision indicate that multi-task learning results in performance benefits. Additional exploration of employing Knowledge Base priors into theVAE reveals that the sentence space can be shifted towards that of the Knowledge Base, offering interpretability and further improving results.

Document-level relation extraction is a complex human process that requires logical inference to extract relationships between named entities in text. Existing approaches use graph-based neural models with words as nodes and edges as relations between them, to encode relations across sentences. These models are node-based, i.e., they form pair representations based solely on the two target node representations. However, entity relations can be better expressed through unique edge representations formed as paths between nodes. We thus propose an edge-oriented graph neural model for document-level relation extraction. The model utilises different types of nodes and edges to create a document-level graph. An inference mechanism on the graph edges enables to learn intra- and inter-sentence relations using multi-instance learning internally. Experiments on two document-level biomedical datasets for chemical-disease and gene-disease associations show the usefulness of the proposed edge-oriented approach.