What was the topic of your PhD research and why was it an interesting area? My topic of research was developing an optical tactile sensor to track head motion during radiotherapy. I worked on both the hardware and software development of this sensor, though my focus was mostly on the software side. Its importance comes from the fact that during radiotherapy, patients undergoing head and neck cancer treatment are typically immobilised. This is usually done using a thermoplastic mask, which can feel very claustrophobic, or a stereotactic frame.

We appreciate the reviewers for your1 commendation for the simplicity, intuitiveness and effectiveness of our method. About theoretical analysis: The main contribution of our paper is a novel early exiting approach that empirically4 performs well. For DeeBERT, we use the official code to obtain the resultson the development set.

Thank reviewers for detailed comments. Our main contribution is the novel search-and-learning for1 UnsupTextGen, achieving remarkable performance (sometimes even better than SupTextGen). Forexample,8 the rules/heuristics in [18] only applies to sentiment style transfer.Future work MT: Thanks for9 suggesting the future work. We are currently considering MT by using word-level dictionary and10 performingsearchandlearning. R2: 3(Novelty): Our main novelty is the search-and-learning framework TGSL for UnsupTextGen,16 where our learning is non-trivial and involves two stages with different losses, well motivated and17 supportedbyablationstudy.

Cryptic crosswords, the dominant crossword variety in the UK, are a promising target for advancing NLP systems that seek to process semantically complex, highly compositional language. Cryptic clues read like fluent natural language but are adversarially composed of two parts: a definition and a wordplay cipher requiring character-level manipulations. Expert humans use creative intelligence to solve cryptics, flexibly combining linguistic, world, and domain knowledge. In this paper, we make two main contributions. First, we present a dataset of cryptic clues as a challenging new benchmark for NLP systems that seek to process compositional language in more creative, human-like ways. After showing that three non-neural approaches and T5, a state-of-the-art neural language model, do not achieve good performance, we make our second main contribution: a novel curriculum approach, in which the model is first fine-tuned on related tasks such as unscrambling words. We also introduce a challenging data split, examine the meta-linguistic capabilities of subword-tokenized models, and investigate model systematicity by perturbing the wordplay part of clues, showing that T5 exhibits behavior partially consistent with human solving strategies. Although our curricular approach considerably improves on the T5 baseline, our best-performing model still fails to generalize to the extent that humans can. Thus, cryptic crosswords remain an unsolved challenge for NLP systems and a potential source of future innovation.