Finetuning language agents with reasoning-action trajectories is effective, but obtaining these trajectories from human annotations or stronger models is costly and sometimes impractical. In this paper, we investigate the use of self-training in language agents, which can generate supervision from the agent itself, offering a promising alternative without relying on human or stronger model demonstrations. Self-training, however, requires high-quality model-generated samples, which are hard to obtain for challenging language agent tasks. To address this, we present Reflection-Reinforced Self-Training (Re-ReST), which uses a reflector to refine low-quality generated samples during self-training. The reflector takes the agent’s output and feedback from an external environment (e.g., unit test results in code generation) to produce improved samples. This technique enhances the quality of inferior samples and efficiently enriches the self-training dataset with higher-quality samples. We conduct extensive experiments on open-source language agents across tasks, including multi-hop question answering, sequential decision-making, code generation, visual question answering, and text-to-image generation. The results demonstrate the effectiveness of self-training and Re-ReST in language agent tasks, with self-training improving baselines by 7.6% on HotpotQA and 28.4% on AlfWorld, and Re-ReST further boosting performance by 2.0% and 14.1%, respectively. Our studies also confirm the efficiency of using a reflector to generate high-quality samples for self-training. Moreover, we demonstrate a method to employ reflection during inference without ground-truth feedback, addressing the limitation of previous reflection work.

Large Vision-Language Models (LVLMs) suffer from hallucination issues, wherein the models generate plausible-sounding but factually incorrect outputs, undermining their reliability. A comprehensive quantitative evaluation is necessary to identify and understand the extent of hallucinations in these models. However, existing benchmarks are often limited in scope, focusing mainly on object hallucinations. Furthermore, current evaluation methods struggle to effectively address the subtle semantic distinctions between model outputs and reference data, as well as the balance between hallucination and informativeness. To address these issues, we introduce a multi-dimensional benchmark covering objects, attributes, and relations, with challenging images selected based on associative biases. Moreover, we propose a large language model (LLM)-based two-stage evaluation framework that generalizes the popular CHAIR metric and incorporates both faithfulness and coverage into the evaluation. Experiments on 10 established LVLMs demonstrate that our evaluation metric is more comprehensive and better correlated with humans than existing work when evaluating on our challenging human-annotated benchmark dataset. Our work also highlights the critical balance between faithfulness and coverage of model outputs, and encourages future works to address hallucinations in LVLMs while keeping their outputs informative.

There is increasing interest in equipping language models with the ability to leverage external tools for complex, goal-oriented tasks. However, interacting with external tools introduces inherent uncertainties due to imperfections and misalignments between the tools’ outputs and the agents’ internal models, often leading to suboptimal outcomes. We thus study the problem of tool-use calibration in language agents, and identify prompt design and execution trace selection as two primary areas that suffer from miscalibration. We then propose ProbeCal, which recalibrates the internal probabilities of tool-using language agents to better reflect the actual effectiveness of tool, and enables a more appropriate selection of prompts and execution paths. We empirically show that ProbeCal can significantly and consistently improve off-the-shelf language models in tool-using applications.

Vision-language models have become increasingly powerful for tasks that require an understanding of both visual and linguistic elements, bridging the gap between these modalities. In the context of multimodal clinical AI, there is a growing need for models that possess domain-specific knowledge, as existing models often lack the expertise required for medical applications. In this paper, we take brain abnormalities as an example to demonstrate how to automatically collect medical image-text aligned data for pretraining from public resources such as PubMed. In particular, we present a pipeline that streamlines the pre-training process by initially collecting a large brain image-text dataset from case reports and published journals and subsequently constructing a high-performance vision-language model tailored to specific medical tasks. We also investigate the unique challenge of mapping subfigures to subcaptions in the medical domain. We evaluated the resulting model with quantitative and qualitative intrinsic evaluations. The resulting dataset will be released to the community.

Vision-and-language navigation (VLN) agents are trained to navigate in real-world environments based on natural language instructions. A major challenge in VLN is the limited available training data, which hinders the models’ ability to generalize effectively. Previous approaches have attempted to alleviate this issue by using external tools to generate pseudo-labeled data or integrating web-scaled image-text pairs during training. However, these methods often rely on automatically-generated or out-of-domain data, leading to challenges such as suboptimal data quality and domain mismatch. In this paper, we introduce a masked path modeling (MPM) objective. MPM pretrains an agent using self-collected data for subsequent navigation tasks, eliminating the need for external tools. Specifically, our method allows the agent to explore navigation environments and record the paths it traverses alongside the corresponding agent actions. Subsequently, we train the agent on this collected data to reconstruct the original action sequence when given a randomly masked subsequence of the original path. This approach enables the agent to accumulate a diverse and substantial dataset, facilitating the connection between visual observations of paths and the agent’s actions, which is the foundation of the VLN task. Importantly, the collected data are in-domain, and the training process avoids synthetic data with uncertain quality, addressing previous issues. We conduct experiments on various VLN datasets and demonstrate the applications of MPM across different levels of instruction complexity. Our results exhibit significant improvements in success rates, with enhancements of 1.3%, 1.1%, and 1.2% on the val-unseen split of the Room-to-Room, Room-for-Room, and Room-across-Room datasets, respectively. Additionally, we underscore the adaptability of MPM as well as the potential for additional improvements when the agent is allowed to explore unseen environments prior to testing.

Model-based evaluation metrics (e.g., CLIPScore and GPTScore) have demonstrated decent correlations with human judgments in various language generation tasks. However, their impact on fairness remains largely unexplored. It is widely recognized that pretrained models can inadvertently encode societal biases, thus employing these models for evaluation purposes may inadvertently perpetuate and amplify biases. For example, an evaluation metric may favor the caption “a woman is calculating an account book” over “a man is calculating an account book,” even if the image only shows male accountants. In this paper, we conduct a systematic study of gender biases in model-based automatic evaluation metrics for image captioning tasks. We start by curating a dataset comprising profession, activity, and object concepts associated with stereotypical gender associations. Then, we demonstrate the negative consequences of using these biased metrics, including the inability to differentiate between biased and unbiased generations, as well as the propagation of biases to generation models through reinforcement learning. Finally, we present a simple and effective way to mitigate the metric bias without hurting the correlations with human judgments. Our dataset and framework lay the foundation for understanding the potential harm of model-based evaluation metrics, and facilitate future works to develop more inclusive evaluation metrics.

Multimodal counterfactual reasoning is a vital yet challenging ability for AI systems. It involves predicting the outcomes of hypothetical circumstances based on vision and language inputs, which enables AI models to learn from failures and explore hypothetical scenarios. Despite its importance, there are only a few datasets targeting the counterfactual reasoning abilities of multimodal models. Among them, they only cover reasoning over synthetic environments or specific types of events (e.g. traffic collisions), making them hard to reliably benchmark the model generalization ability in diverse real-world scenarios and reasoning dimensions. To overcome these limitations, we develop a video question answering dataset, ACQUIRED: it consists of 3.9K annotated videos, encompassing a wide range of event types and incorporating both first and third-person viewpoints, which ensures a focus on real-world diversity. In addition, each video is annotated with questions that span three distinct dimensions of reasoning, including physical, social, and temporal, which can comprehensively evaluate the model counterfactual abilities along multiple aspects. We benchmark our dataset against several state-of-the-art language-only and multimodal models and experimental results demonstrate a significant performance gap (>13%) between models and humans. The findings suggest that multimodal counterfactual reasoning remains an open challenge and ACQUIRED is a comprehensive and reliable benchmark for inspiring future research in this direction.

The speaker-follower models have proven to be effective in vision-and-language navigation, where a speaker model is used to synthesize new instructions to augment the training data for a follower navigation model. However, in previous work, the speaker model is follower-agnostic and fails to take the state of the follower into consideration. In this paper, we present FOAM, a FOllower-Aware speaker Model that is constantly updated given the follower feedback, so that the generated instructions can be more suitable to the current learning state of the follower. Specifically, we optimize the speaker using a bi-level optimization framework and obtain its training signals by evaluating the follower on labeled data. Experimental results on the Room-to-Room and Room-across-Room datasets demonstrate that our methods can outperform strong baseline models across settings. Analyses also reveal that our generated instructions are of higher quality than the baselines.

Although some recent works show potential complementarity among different state-of-the-art systems, few works try to investigate this problem in text summarization. Researchers in other areas commonly refer to the techniques of reranking or stacking to approach this problem. In this work, we highlight several limitations of previous methods, which motivates us to present a new framework Refactor that provides a unified view of text summarization and summaries combination. Experimentally, we perform a comprehensive evaluation that involves twenty-two base systems, four datasets, and three different application scenarios. Besides new state-of-the-art results on CNN/DailyMail dataset (46.18 ROUGE-1), we also elaborate on how our proposed method addresses the limitations of the traditional methods and the effectiveness of the Refactor model sheds light on insight for performance improvement. Our system can be directly used by other researchers as an off-the-shelf tool to achieve further performance improvements. We open-source all the code and provide a convenient interface to use it: https://github.com/yixinL7/Refactoring-Summarization.

Neural abstractive summarization models are flexible and can produce coherent summaries, but they are sometimes unfaithful and can be difficult to control. While previous studies attempt to provide different types of guidance to control the output and increase faithfulness, it is not clear how these strategies compare and contrast to each other. In this paper, we propose a general and extensible guided summarization framework (GSum) that can effectively take different kinds of external guidance as input, and we perform experiments across several different varieties. Experiments demonstrate that this model is effective, achieving state-of-the-art performance according to ROUGE on 4 popular summarization datasets when using highlighted sentences as guidance. In addition, we show that our guided model can generate more faithful summaries and demonstrate how different types of guidance generate qualitatively different summaries, lending a degree of controllability to the learned models.

Word alignment over parallel corpora has a wide variety of applications, including learning translation lexicons, cross-lingual transfer of language processing tools, and automatic evaluation or analysis of translation outputs. The great majority of past work on word alignment has worked by performing unsupervised learning on parallel text. Recently, however, other work has demonstrated that pre-trained contextualized word embeddings derived from multilingually trained language models (LMs) prove an attractive alternative, achieving competitive results on the word alignment task even in the absence of explicit training on parallel data. In this paper, we examine methods to marry the two approaches: leveraging pre-trained LMs but fine-tuning them on parallel text with objectives designed to improve alignment quality, and proposing methods to effectively extract alignments from these fine-tuned models. We perform experiments on five language pairs and demonstrate that our model can consistently outperform previous state-of-the-art models of all varieties. In addition, we demonstrate that we are able to train multilingual word aligners that can obtain robust performance on different language pairs.

Phrase grounding aims to map textual phrases to their associated image regions, which can be a prerequisite for multimodal reasoning and can benefit tasks requiring identifying objects based on language. With pre-trained vision-and-language models achieving impressive performance across tasks, it remains unclear if we can directly utilize their learned embeddings for phrase grounding without fine-tuning. To this end, we propose a method to extract matched phrase-region pairs from pre-trained vision-and-language embeddings and propose four fine-tuning objectives to improve the model phrase grounding ability using image-caption data without any supervised grounding signals. Experiments on two representative datasets demonstrate the effectiveness of our objectives, outperforming baseline models in both weakly-supervised and supervised phrase grounding settings. In addition, we evaluate the aligned embeddings on several other downstream tasks and show that we can achieve better phrase grounding without sacrificing representation generality.

Cross-lingual text summarization aims at generating a document summary in one language given input in another language. It is a practically important but under-explored task, primarily due to the dearth of available data. Existing methods resort to machine translation to synthesize training data, but such pipeline approaches suffer from error propagation. In this work, we propose an end-to-end cross-lingual text summarization model. The model uses reinforcement learning to directly optimize a bilingual semantic similarity metric between the summaries generated in a target language and gold summaries in a source language. We also introduce techniques to pre-train the model leveraging monolingual summarization and machine translation objectives. Experimental results in both English–Chinese and English–German cross-lingual summarization settings demonstrate the effectiveness of our methods. In addition, we find that reinforcement learning models with bilingual semantic similarity as rewards generate more fluent sentences than strong baselines.

The COVID-19 pandemic is the worst pandemic to strike the world in over a century. Crucial to stemming the tide of the SARS-CoV-2 virus is communicating to vulnerable populations the means by which they can protect themselves. To this end, the collaborators forming the Translation Initiative for COvid-19 (TICO-19) have made test and development data available to AI and MT researchers in 35 different languages in order to foster the development of tools and resources for improving access to information about COVID-19 in these languages. In addition to 9 high-resourced, ”pivot” languages, the team is targeting 26 lesser resourced languages, in particular languages of Africa, South Asia and South-East Asia, whose populations may be the most vulnerable to the spread of the virus. The same data is translated into all of the languages represented, meaning that testing or development can be done for any pairing of languages in the set. Further, the team is converting the test and development data into translation memories (TMXs) that can be used by localizers from and to any of the languages.

Neural network-based models augmented with unsupervised pre-trained knowledge have achieved impressive performance on text summarization. However, most existing evaluation methods are limited to an in-domain setting, where summarizers are trained and evaluated on the same dataset. We argue that this approach can narrow our understanding of the generalization ability for different summarization systems. In this paper, we perform an in-depth analysis of characteristics of different datasets and investigate the performance of different summarization models under a cross-dataset setting, in which a summarizer trained on one corpus will be evaluated on a range of out-of-domain corpora. A comprehensive study of 11 representative summarization systems on 5 datasets from different domains reveals the effect of model architectures and generation ways (i.e. abstractive and extractive) on model generalization ability. Further, experimental results shed light on the limitations of existing summarizers. Brief introduction and supplementary code can be found in https://github.com/zide05/CDEvalSumm.

Back-translation has proven to be an effective method to utilize monolingual data in neural machine translation (NMT), and iteratively conducting back-translation can further improve the model performance. Selecting which monolingual data to back-translate is crucial, as we require that the resulting synthetic data are of high quality and reflect the target domain. To achieve these two goals, data selection and weighting strategies have been proposed, with a common practice being to select samples close to the target domain but also dissimilar to the average general-domain text. In this paper, we provide insights into this commonly used approach and generalize it to a dynamic curriculum learning strategy, which is applied to iterative back-translation models. In addition, we propose weighting strategies based on both the current quality of the sentence and its improvement over the previous iteration. We evaluate our models on domain adaptation, low-resource, and high-resource MT settings and on two language pairs. Experimental results demonstrate that our methods achieve improvements of up to 1.8 BLEU points over competitive baselines.

Multi-head attention is appealing for its ability to jointly extract different types of information from multiple representation subspaces. Concerning the information aggregation, a common practice is to use a concatenation followed by a linear transformation, which may not fully exploit the expressiveness of multi-head attention. In this work, we propose to improve the information aggregation for multi-head attention with a more powerful routing-by-agreement algorithm. Specifically, the routing algorithm iteratively updates the proportion of how much a part (i.e. the distinct information learned from a specific subspace) should be assigned to a whole (i.e. the final output representation), based on the agreement between parts and wholes. Experimental results on linguistic probing tasks and machine translation tasks prove the superiority of the advanced information aggregation over the standard linear transformation.

In this paper, we describe compare-mt, a tool for holistic analysis and comparison of the results of systems for language generation tasks such as machine translation. The main goal of the tool is to give the user a high-level and coherent view of the salient differences between systems that can then be used to guide further analysis or system improvement. It implements a number of tools to do so, such as analysis of accuracy of generation of particular types of words, bucketed histograms of sentence accuracies or counts based on salient characteristics, and extraction of characteristic n-grams for each system. It also has a number of advanced features such as use of linguistic labels, source side data, or comparison of log likelihoods for probabilistic models, and also aims to be easily extensible by users to new types of analysis. compare-mt is a pure-Python open source package, that has already proven useful to generate analyses that have been used in our published papers. Demo Video: https://youtu.be/NyJEQT7t2CA

Learning general representations of text is a fundamental problem for many natural language understanding (NLU) tasks. Previously, researchers have proposed to use language model pre-training and multi-task learning to learn robust representations. However, these methods can achieve sub-optimal performance in low-resource scenarios. Inspired by the recent success of optimization-based meta-learning algorithms, in this paper, we explore the model-agnostic meta-learning algorithm (MAML) and its variants for low-resource NLU tasks. We validate our methods on the GLUE benchmark and show that our proposed models can outperform several strong baselines. We further empirically demonstrate that the learned representations can be adapted to new tasks efficiently and effectively.

The recent success of neural machine translation models relies on the availability of high quality, in-domain data. Domain adaptation is required when domain-specific data is scarce or nonexistent. Previous unsupervised domain adaptation strategies include training the model with in-domain copied monolingual or back-translated data. However, these methods use generic representations for text regardless of domain shift, which makes it infeasible for translation models to control outputs conditional on a specific domain. In this work, we propose an approach that adapts models with domain-aware feature embeddings, which are learned via an auxiliary language modeling task. Our approach allows the model to assign domain-specific representations to words and output sentences in the desired domain. Our empirical results demonstrate the effectiveness of the proposed strategy, achieving consistent improvements in multiple experimental settings. In addition, we show that combining our method with back translation can further improve the performance of the model.

Neural networks are known to be data hungry and domain sensitive, but it is nearly impossible to obtain large quantities of labeled data for every domain we are interested in. This necessitates the use of domain adaptation strategies. One common strategy encourages generalization by aligning the global distribution statistics between source and target domains, but one drawback is that the statistics of different domains or tasks are inherently divergent, and smoothing over these differences can lead to sub-optimal performance. In this paper, we propose the framework of Domain Differential Adaptation (DDA), where instead of smoothing over these differences we embrace them, directly modeling the difference between domains using models in a related task. We then use these learned domain differentials to adapt models for the target task accordingly. Experimental results on domain adaptation for neural machine translation demonstrate the effectiveness of this strategy, achieving consistent improvements over other alternative adaptation strategies in multiple experimental settings.

Bilingual lexicon extraction has been studied for decades and most previous methods have relied on parallel corpora or bilingual dictionaries. Recent studies have shown that it is possible to build a bilingual dictionary by aligning monolingual word embedding spaces in an unsupervised way. With the recent advances in generative models, we propose a novel approach which builds cross-lingual dictionaries via latent variable models and adversarial training with no parallel corpora. To demonstrate the effectiveness of our approach, we evaluate our approach on several language pairs and the experimental results show that our model could achieve competitive and even superior performance compared with several state-of-the-art models.

Advanced neural machine translation (NMT) models generally implement encoder and decoder as multiple layers, which allows systems to model complex functions and capture complicated linguistic structures. However, only the top layers of encoder and decoder are leveraged in the subsequent process, which misses the opportunity to exploit the useful information embedded in other layers. In this work, we propose to simultaneously expose all of these signals with layer aggregation and multi-layer attention mechanisms. In addition, we introduce an auxiliary regularization term to encourage different layers to capture diverse information. Experimental results on widely-used WMT14 English-German and WMT17 Chinese-English translation data demonstrate the effectiveness and universality of the proposed approach.

Document-level sentiment classification is a fundamental problem which aims to predict a user’s overall sentiment about a product in a document. Several methods have been proposed to tackle the problem whereas most of them fail to consider the influence of users who express the sentiment and products which are evaluated. To address the issue, we propose a deep memory network for document-level sentiment classification which could capture the user and product information at the same time. To prove the effectiveness of our algorithm, we conduct experiments on IMDB and Yelp datasets and the results indicate that our model can achieve better performance than several existing methods.