As more data, better algorithms, and higher computing power continue to shape the future of artificial intelligence (AI), reliable machine learning …

This is the concluding part of our blog where we focus on extracting actionable items from stand-up meetings. You can find the previous blog post here. Just to summarize we looked at how GPT-3 summarized standup calls without any text pre-processing. The summary was good but there was nothing actionable that we could extract because " Who did what " details are important to make something actionable. So in order to improve results, lets think of a standup meeting flow in terms of a block diagram.

Integrating external knowledge into commonsense reasoning tasks has shown progress in resolving some, but not all, knowledge gaps in these tasks. For knowledge integration to yield peak performance, it is critical to select a knowledge graph (KG) that is well-aligned with the given task's objective. We present an approach to assess how well a candidate KG can correctly identify and accurately fill in gaps of reasoning for a task, which we call KG-to-task match. We show this KG-to-task match in 3 phases: knowledge-task identification, knowledge-task alignment, and knowledge-task integration. We also analyze our transformer-based KG-to-task models via commonsense probes to measure how much knowledge is captured in these models before and after KG integration. Empirically, we investigate KG matches for the SocialIQA (SIQA) (Sap et al., 2019b), Physical IQA (PIQA) (Bisk et al., 2020), and MCScript2.0 (Ostermann et al., 2019) datasets with 3 diverse KGs: ATOMIC (Sap et al., 2019a), ConceptNet (Speer et al., 2017), and an automatically constructed instructional KG based on WikiHow (Koupaee and Wang, 2018). With our methods we are able to demonstrate that ATOMIC, an event-inference focused KG, is the best match for SIQA and MCScript2.0, and that the taxonomic ConceptNet and WikiHow-based KGs are the best matches for PIQA across all 3 analysis phases. We verify our methods and findings with human evaluation.

Position encoding in transformer architecture provides supervision for dependency modeling between elements at different positions in the sequence. We investigate various methods to encode positional information in transformer-based language models and propose a novel implementation named Rotary Position Embedding(RoPE). The proposed RoPE encodes absolute positional information with rotation matrix and naturally incorporates explicit relative position dependency in self-attention formulation. Notably, RoPE comes with valuable properties such as flexibility of being expand to any sequence lengths, decaying inter-token dependency with increasing relative distances, and capability of equipping the linear self-attention with relative position encoding. As a result, the enhanced transformer with rotary position embedding, or RoFormer, achieves superior performance in tasks with long texts. We release the theoretical analysis along with some preliminary experiment results on Chinese data. The undergoing experiment for English benchmark will soon be updated.

Large pretrained generative models like GPT-3 often suffer from hallucinating non-existent or incorrect content, which undermines their potential merits in real applications. Existing work usually attempts to detect these hallucinations based on a corresponding oracle reference at a sentence or document level. However ground-truth references may not be readily available for many free-form text generation applications, and sentence- or document-level detection may fail to provide the fine-grained signals that would prevent fallacious content in real time. As a first step to addressing these issues, we propose a novel token-level, reference-free hallucination detection task and an associated annotated dataset named HaDes (HAllucination DEtection dataSet). To create this dataset, we first perturb a large number of text segments extracted from English language Wikipedia, and then verify these with crowd-sourced annotations. To mitigate label imbalance during annotation, we utilize an iterative model-in-loop strategy. We conduct comprehensive data analyses and create multiple baseline models.

Large-scale language models such as GPT-3 are excellent few-shot learners, allowing them to be controlled via natural text prompts. Recent studies report that prompt-based direct classification eliminates the need for fine-tuning but lacks data and inference scalability. This paper proposes a novel data augmentation technique that leverages large-scale language models to generate realistic text samples from a mixture of real samples. We also propose utilizing soft-labels predicted by the language models, effectively distilling knowledge from the large-scale language models and creating textual perturbations simultaneously. We perform data augmentation experiments on diverse classification tasks and show that our method hugely outperforms existing text augmentation methods. Ablation studies and a qualitative analysis provide more insights into our approach.

When primed with only a handful of training samples, very large pretrained language models such as GPT-3, have shown competitive results when compared to fully-supervised fine-tuned large pretrained language models. We demonstrate that the order in which the samples are provided can be the difference between near state-of-the-art and random guess performance: Essentially some permutations are "fantastic" and some not. We analyse this phenomenon in detail, establishing that: it is present across model sizes (even for the largest current models), it is not related to a specific subset of samples, and that a given good permutation for one model is not transferable to another. While one could use a development set to determine which permutations are performant, this would deviate from the few-shot setting as it requires additional annotated data. Instead, we use the generative nature of the language models to construct an artificial development set and based on entropy statistics of the candidate permutations from this set we identify performant prompts. Our method improves upon GPT-family models by on average 13% relative across eleven different established text classification tasks.

Can we enable NLP models to appropriately respond to instructional prompts and consequently generalize to new tasks? To study this question, we leverage the existing NLP datasets and the instructions that were used to crowdsource them to create NATURAL INSTRUCTIONS, a dataset of instructions and task-specific input/output data. This dataset consists of 61 distinct language instructions and about 600k task instances, and is used to evaluate existing state-of-the-art language-models (LMs) in addressing new tasks by few-shot prompting of GPT3 and fine-tuning BART. Our analysis indicates that: (a) the existing models indeed benefit from instructions and hence, show improved generalization to new tasks; (b) while models like GPT-3 generally benefit from instructions, the extent of their gains varies across different fields of instructions and also depends on the task being solved; (c) generalization to unseen tasks in NATURAL INSTRUCTIONS remains far from perfect for the state-of-the-art, indicating significant room for more progress in this direction.

As language models are trained on ever more text, researchers are turning to some of the largest corpora available. Unlike most other types of datasets in NLP, large unlabeled text corpora are often presented with minimal documentation, and best practices for documenting them have not been established. In this work we provide the first documentation for the Colossal Clean Crawled Corpus (C4; Raffel et al., 2020), a dataset created by applying a set of filters to a single snapshot of Common Crawl. We begin with a high-level summary of the data, including distributions of where the text came from and when it was written. We then give more detailed analysis on salient parts of this data, including the most frequent sources of text (e.g., patents.google.com, which contains a significant percentage of machine translated and/or OCR'd text), the effect that the filters had on the data (they disproportionately remove text in AAE), and evidence that some other benchmark NLP dataset examples are contained in the text. We release a web interface to an interactive, indexed copy of this dataset, encouraging the community to continuously explore and report additional findings.

The idea of artificial general intelligence as we know it today starts with a dot-com blowout on Broadway. Twenty years ago--before Shane Legg clicked with neuroscience postgrad Demis Hassabis over a shared fascination with intelligence; before the pair hooked up with Hassabis's childhood friend Mustafa Suleyman, a progressive activist, to spin that fascination into a company called DeepMind; before Google bought that company for more than half a billion dollars four years later--Legg worked at a startup in New York called Webmind, set up by AI researcher Ben Goertzel. Today the two men represent two very different branches of the future of artificial intelligence, but their roots reach back to common ground. Even for the heady days of the dot-com bubble, Webmind's goals were ambitious. Goertzel wanted to create a digital baby brain and release it onto the internet, where he believed it would grow up to become fully self-aware and far smarter than humans.