No abstract

Plug-and-play language models (PPLMs) enable topic-conditioned natural language generation by combining large pre-trained generators with attribute models to steer the predicted token distribution towards selected topics. Despite their efficiency, the large amounts of labeled texts required by PPLMs to effectively balance generation fluency and proper conditioning make them unsuitable to low-resource scenarios. We present ETC-NLG, an approach leveraging topic modeling annotations to produce End-to-end Topic-Conditioned Natural Language Generations over emergent topics in unlabeled document collections. We test our method's effectiveness in a low-resource setting for Italian and perform a comparative evaluation of ETC-NLG for Italian and English using a parallel corpus. Finally, we propose an evaluation method to automatically estimate the conditioning effectiveness from generated utterances.

No abstract

Plug-and-play language models (PPLMs) enable topic-conditioned natural language generation by pairing large pre-trained generators with attribute models used to steer the predicted token distribution towards the selected topic. Despite their computational efficiency, PPLMs require large amounts of labeled texts to effectively balance generation fluency and proper conditioning, making them unsuitable for low-resource settings. We present ETC-NLG, an approach leveraging topic modeling annotations to enable fully-unsupervised End-toend Topic-Conditioned Natural Language Generation over emergent topics in unlabeled document collections. We first test the effectiveness of our approach in a lowresource setting for Italian, evaluating the conditioning for both topic models and gold annotations. We then perform a comparative evaluation of ETC-NLG for Italian and English using a parallel corpus. Finally, we propose an automatic approach to estimate the effectiveness of conditioning on generated utterances.

Deep Learning models for protein structure prediction, such as AlphaFold2, leverage Transformer architectures and their attention mechanism to capture structural and functional properties of amino acid sequences. Despite the high accuracy of predictions, biologically insignificant perturbations of the input sequences, or even single point mutations, can lead to substantially different 3d structures. On the other hand, protein language models are often insensitive to biologically relevant mutations that induce misfolding or dysfunction (e.g. missense mutations). Precisely, predictions of the 3d coordinates do not reveal the structure-disruptive effect of these mutations. Therefore, there is an evident inconsistency between the biological importance of mutations and the resulting change in structural prediction. Inspired by this problem, we introduce the concept of adversarial perturbation of protein sequences in continuous embedding spaces of protein language models. Our method relies on attention scores to detect the most vulnerable amino acid positions in the input sequences. Adversarial mutations are biologically diverse from their references and are able to significantly alter the resulting 3d structures.