Aspect-Sentiment Triplet Extraction (ASTE) is a recently proposed task of aspect-based sentiment analysis that consists in extracting (aspect phrase, opinion phrase, sentiment polarity) triples from a given sentence. Recent state-of-the-art methods approach this task by first extracting all possible text spans from a given text, then filtering the potential aspect and opinion phrases with a classifier, and finally considering all their pairs with another classifier that additionally assigns sentiment polarity to them. Although several variations of the above scheme have been proposed, the common feature is that the final result is constructed by a sequence of independent classifier decisions. This hinders the exploitation of dependencies between extracted phrases and prevents the use of knowledge about the interrelationships between classifier predictions to improve performance. In this paper, we propose a new ASTE approach consisting of three transformer-inspired layers, which enables the modelling of dependencies both between phrases and between the final classifier decisions. Experimental results show that the method achieves higher performance in terms of F1 measure than other methods studied on popular benchmarks. In addition, we show that a simple pre-training technique further improves the performance of the model.

In dialogue, the addressee may initially misunderstand the speaker and respond erroneously, often prompting the speaker to correct the misunderstanding in the next turn with a Third Position Repair (TPR). The ability to process and respond appropriately to such repair sequences is thus crucial in conversational AI systems. In this paper, we first collect, analyse, and publicly release BlockWorld-Repairs: a dataset of multi-modal TPR sequences in an instruction-following manipulation task that is, by design, rife with referential ambiguity. We employ this dataset to evaluate several state-of-the-art Vision and Language Models (VLM) across multiple settings, focusing on their capability to process and accurately respond to TPRs and thus recover from miscommunication. We find that, compared to humans, all models significantly underperform in this task. We then show that VLMs can benefit from specialised losses targeting relevant tokens during fine-tuning, achieving better performance and generalising better to new scenarios. Our results suggest that these models are not yet ready to be deployed in multi-modal collaborative settings where repairs are common, and highlight the need to design training regimes and objectives that facilitate learning from interaction. Our code and data are available at www.github.com/JChiyah/blockworld-repairs

In this paper, we investigate the influence of claims in analyst reports and earnings calls on financial market returns, considering them as significant quarterly events for publicly traded companies. To facilitate a comprehensive analysis, we construct a new financial dataset for the claim detection task in the financial domain. We benchmark various language models on this dataset and propose a novel weak-supervision model that incorporates the knowledge of subject matter experts (SMEs) in the aggregation function, outperforming existing approaches. We also demonstrate the practical utility of our proposed model by constructing a novel measure of optimism. Here, we observe the dependence of earnings surprise and return on our optimism measure. Our dataset, models, and code are publicly (under CC BY 4.0 license) available on GitHub.

We propose an Autoregressive (AR) Moving-average (MA) attention structure that can adapt to various linear attention mechanisms, enhancing their ability to capture long-range and local temporal patterns in time series. In this paper, we first demonstrate that, for the time series forecasting (TSF) task, the previously overlooked decoder-only autoregressive Transformer model can achieve results comparable to the best baselines when appropriate tokenization and training methods are applied. Moreover, inspired by the ARMA model from statistics and recent advances in linear attention, we introduce the full ARMA structure into existing autoregressive attention mechanisms. By using an indirect MA weight generation method, we incorporate the MA term while maintaining the time complexity and parameter size of the underlying efficient attention models. We further explore how indirect parameter generation can produce implicit MA weights that align with the modeling requirements for local temporal impacts. Experimental results show that incorporating the ARMA structure consistently improves the performance of various AR attentions on TSF tasks, achieving state-of-the-art results.

Computational methods to aid journalists in the task often require adapting a model to specific domains and generating explanations. However, most automated fact-checking methods rely on three-class datasets, which do not accurately reflect real-world misinformation. Moreover, fact-checking explanations are often generated based on text summarization of evidence, failing to address the relationship between the claim and the evidence. To address these issues, we extend the self-rationalization method--typically used in natural language inference (NLI) tasks--to fact verification. We propose a label-adaptive learning approach: first, we fine-tune a model to learn veracity prediction with annotated labels (step-1 model). Then, we fine-tune the step-1 model again to learn self-rationalization, using the same data and additional annotated explanations. Our results show that our label-adaptive approach improves veracity prediction by more than ten percentage points (Macro F1) on both the PubHealth and AVeriTec datasets, outperforming the GPT-4 model. Furthermore, to address the high cost of explanation annotation, we generated 64 synthetic explanations from three large language models: GPT-4-turbo, GPT-3.5-turbo, and Llama-3-8B and few-shot fine-tune our step-1 model. The few-shot synthetic explanation fine-tuned model performed comparably to the fully fine-tuned self-rationalization model, demonstrating the potential of low-budget learning with synthetic data. Our label-adaptive self-rationalization approach presents a promising direction for future research on real-world explainable fact-checking with different labeling schemes.

OpenAI is beta testing a new workspace interface for ChatGPT called Canvas. The AI giant unveiled its new ChatGPT workspace on its official blog and it's currently available for ChatGPT Plus and Team users. Enterprise and Edu users will be able to access Canvas sometime next week. Canvas is a virtual interface space for writing and coding projects that allow users to consult with ChatGPT on certain portions of a project. A separate window opens besides the main chat space and users can put writing or code on this new "canvas" and highlight sections to have the model focus on and edit "like a copy editor or code reviewer," according to the blog.

Just one day after OpenAI announced a 6.6 billion funding round, the company is launching its first major interface evolution for ChatGPT. In what could be recognition from OpenAI that its transformational chatbot is ready for user experiences beyond a question and answer format, the new beta feature is an editable canvas that opens in a window alongside ChatGPT's standard chatbox. "The core thing we're trying to solve is a better way to collaborate with ChatGPT on writing and coding," says Daniel Levine, a product lead at OpenAI for the canvas feature. Canvas is rolling out in beta to ChatGPT Plus and Team subscribers today, and Enterprise and Edu customers will likely get the feature next week. The feature is fully functional on desktops--mobile users can only view the canvas projects for now.

Keeping ChatGPT running is expensive as heck, so OpenAI needs access to plenty of cash to make sure the lights stay on. A day after the company said it had secured 6.6 billion in funding -- the biggest ever funding round for a startup -- it confirmed that it has a new 4 billion revolving line of credit. OpenAI has yet to tap the credit line, which it obtained from JPMorgan Chase, Citi, Goldman Sachs, Morgan Stanley, Santander, Wells Fargo, SMBC, UBS and HSBC. Some of those banks are also among OpenAI's customers. All told, OpenAI now has a war chest of over 10 billion in liquid funds.

Google has released a new feature which will allow people to search the internet by taking a video. Video search will let people point their camera at something, ask a question about it, and get search results. Android and iPhone users globally will gain access to the feature from 1700 GMT by enabling "AI Overviews" in their Google app, but it will only support English at launch. It is the latest move from the tech giant to change how people search online by utilising artificial intelligence (AI). It comes three months after ChatGPT-maker OpenAI announced it was trialling the ability to search by asking its chatbot questions.

It's a Wednesday afternoon and I've just spent the past 15 minutes texting with a 60-year-old, AI-generated version of myself. My AI future self, which was trained on survey questions I filled out moments before, has just finished spamming me with a string of messages advising me to "stay true" to myself and follow my passions. My 60-year-old AI doppleganger described a fulfilled if slightly boring life. But things suddenly took a turn when I probed the AI about its biggest regrets. After a brief pause, the AI spits out another message explaining how my professional ambitions had led me to neglect my mother in favor of completing my first book.