Though end-to-end neural approaches have recently been dominating NLP tasks in both performance and ease-of-use, they lack interpretability and robustness. We propose Binder, a training-free neural-symbolic framework that maps the task input to a program, which (1) allows binding a unified API of language model (LM) functionalities to a programming language (e.g., SQL, Python) to extend its grammar coverage and thus tackle more diverse questions, (2) adopts an LM as both the program parser and the underlying model called by the API during execution, and (3) requires only a few in-context exemplar annotations. Specifically, we employ GPT-3 Codex as the LM. In the parsing stage, with only a few in-context exemplars, Codex is able to identify the part of the task input that cannot be answerable by the original programming language, correctly generate API calls to prompt Codex to solve the unanswerable part, and identify where to place the API calls while being compatible with the original grammar. In the execution stage, Codex can perform versatile functionalities (e.g., commonsense QA, information extraction) given proper prompts in the API calls. Binder achieves state-of-the-art results on WikiTableQuestions and TabFact datasets, with explicit output programs that benefit human debugging. Note that previous best systems are all finetuned on tens of thousands of task-specific samples, while Binder only uses dozens of annotations as in-context exemplars without any training. Our code is available at https://github.com/HKUNLP/Binder .
We present FOLIO, a human-annotated, open-domain, and logically complex and diverse dataset for reasoning in natural language (NL), equipped with first order logic (FOL) annotations. FOLIO consists of 1,435 examples (unique conclusions), each paired with one of 487 sets of premises which serve as rules to be used to deductively reason for the validity of each conclusion. The logical correctness of premises and conclusions is ensured by their parallel FOL annotations, which are automatically verified by our FOL inference engine. In addition to the main NL reasoning task, NL-FOL pairs in FOLIO automatically constitute a new NL-FOL translation dataset using FOL as the logical form. Our experiments on FOLIO systematically evaluate the FOL reasoning ability of supervised fine-tuning on medium-sized language models (BERT, RoBERTa) and few-shot prompting on large language models (GPT-NeoX, OPT, GPT-3, Codex). For NL-FOL translation, we experiment with GPT-3 and Codex. Our results show that one of the most capable Large Language Model (LLM) publicly available, GPT-3 davinci, achieves only slightly better than random results with few-shot prompting on a subset of FOLIO, and the model is especially bad at predicting the correct truth values for False and Unknown conclusions. Our dataset and code are available at https://github.com/Yale-LILY/FOLIO.
We introduce RealTime QA, a dynamic question answering (QA) platform that announces questions and evaluates systems on a regular basis (weekly in this version). RealTime QA inquires about the current world, and QA systems need to answer questions about novel events or information. It therefore challenges static, conventional assumptions in open domain QA datasets and pursues, instantaneous applications. We build strong baseline models upon large pretrained language models, including GPT-3 and T5. Our benchmark is an ongoing effort, and this preliminary report presents real-time evaluation results over the past month. Our experimental results show that GPT-3 can often properly update its generation results, based on newly-retrieved documents, highlighting the importance of up-to-date information retrieval. Nonetheless, we find that GPT-3 tends to return outdated answers when retrieved documents do not provide sufficient information to find an answer. This suggests an important avenue for future research: can an open domain QA system identify such unanswerable cases and communicate with the user or even the retrieval module to modify the retrieval results? We hope that RealTime QA will spur progress in instantaneous applications of question answering and beyond.
Evaluation in machine learning is usually informed by past choices, for example which datasets or metrics to use. This standardization enables the comparison on equal footing using leaderboards, but the evaluation choices become sub-optimal as better alternatives arise. This problem is especially pertinent in natural language generation which requires ever-improving suites of datasets, metrics, and human evaluation to make definitive claims. To make following best model evaluation practices easier, we introduce GEMv2. The new version of the Generation, Evaluation, and Metrics Benchmark introduces a modular infrastructure for dataset, model, and metric developers to benefit from each others work. GEMv2 supports 40 documented datasets in 51 languages. Models for all datasets can be evaluated online and our interactive data card creation and rendering tools make it easier to add new datasets to the living benchmark.
Program synthesis aims to generate executable programs that are consistent with the user specification. While there are often multiple programs that satisfy the same user specification, existing neural program synthesis models are often only learned from one reference program by maximizing its log-likelihood. This causes the model to be overly confident in its predictions as it sees the single solution repeatedly during training. This leads to poor generalization on unseen examples, even when multiple attempts are allowed. To mitigate this issue, we propose to let the model perform sampling during training and learn from both self-sampled fully-correct programs, which yield the gold execution results, as well as partially-correct programs, whose intermediate execution state matches another correct program. We show that our use of self-sampled correct and partially-correct programs can benefit learning and help guide the sampling process, leading to more efficient exploration of the program space. Additionally, we explore various training objectives to support learning from multiple programs per example and find they greatly affect the performance. Experiments on the MathQA and GSM8K datasets show that our proposed method improves the pass@k performance by 3.1% to 12.3% compared to learning from a single reference program with MLE.
Despite the successes of neural attention models for natural language generation tasks, the quadratic memory complexity of the self-attention module with respect to the input length hinders their applications in long text summarization. Instead of designing more efficient attention modules, we approach this problem by investigating if models with a restricted context can have competitive performance compared with the memory-efficient attention models that maintain a global context by treating the input as an entire sequence. Our model is applied to individual pages, which contain parts of inputs grouped by the principle of locality, during both encoding and decoding stages. We empirically investigated three kinds of localities in text summarization at different levels, ranging from sentences to documents. Our experimental results show that our model can have better performance compared with strong baseline models with efficient attention modules, and our analysis provides further insights of our locality-aware modeling strategy.
Unfaithful text generation is a common problem for text generation systems. In the case of Data-to-Text (D2T) systems, the factuality of the generated text is particularly crucial for any real-world applications. We introduce R2D2, a training framework that addresses unfaithful Data-to-Text generation by training a system both as a generator and a faithfulness discriminator with additional replacement detection and unlikelihood learning tasks. To facilitate such training, we propose two methods for sampling unfaithful sentences. We argue that the poor entity retrieval capability of D2T systems is one of the primary sources of unfaithfulness, so in addition to the existing metrics, we further propose NER-based metrics to evaluate the fidelity of D2T generations. Our experimental results show that R2D2 systems could effectively mitigate the unfaithful text generation, and they achieve new state-of-the-art results on FeTaQA, LogicNLG, and ToTTo, all with significant improvements.
Most graph neural networks (GNNs) rely on the message passing paradigm to propagate node features and build interactions. Recent works point out that different graph learning tasks require different ranges of interactions between nodes. To investigate the underlying mechanism, we explore the capacity of GNNs to capture pairwise interactions between nodes under contexts with different complexities, especially for their graph-level and node-level applications in scientific domains like biochemistry and physics. When formulating pairwise interactions, we study two standard graph construction methods in scientific domains, i.e., K-nearest neighbor (KNN) graphs and fully-connected (FC) graphs. Furthermore, we demonstrate that the inductive bias introduced by KNN-graphs and FC-graphs inhibits GNNs from learning the most informative order of interactions. Such a phenomenon is broadly shared by several GNNs for different graph learning tasks and prevents GNNs from reaching the global minimum loss, so we name it a representation bottleneck. To overcome that, we propose a novel graph rewiring approach based on the pairwise interaction strengths to adjust the reception fields of each node dynamically. Extensive experiments in molecular property prediction and dynamic system forecast prove the superiority of our method over state-of-the-art GNN baselines. Besides, this paper provides a reasonable explanation of why subgraphs play a vital role in determining graph properties. The code is available at https://github.com/smiles724/bottleneck.
Natural language generation technology has recently seen remarkable progress with large-scale training, and many natural language applications are now built upon a wide range of generation models. Combining diverse models may lead to further progress, but conventional ensembling (e.g., shallow fusion) requires that they share vocabulary/tokenization schemes. We introduce Twist decoding, a simple and general inference algorithm that generates text while benefiting from diverse models. Our method does not assume the vocabulary, tokenization or even generation order is shared. Our extensive evaluations on machine translation and scientific paper summarization demonstrate that Twist decoding substantially outperforms each model decoded in isolation over various scenarios, including cases where domain-specific and general-purpose models are both available. Twist decoding also consistently outperforms the popular reranking heuristic where output candidates from one model is rescored by another. We hope that our work will encourage researchers and practitioners to examine generation models collectively, not just independently, and to seek out models with complementary strengths to the currently available models.