Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in vision-language understanding. Recently, with the integration of test-time scaling techniques, these models have also shown strong potential in visual reasoning. However, most existing reasoning approaches remain text-level in nature: MLLMs are prompted to explore various combinations of textual tokens via their underlying language model, while the visual input remains fixed throughout the reasoning process. This paradigm limits the model’s ability to fully exploit rich visual information, particularly when dealing with images containing numerous fine-grained elements. In such cases, vision-level reasoning becomes crucial—where models dynamically zoom into specific regions of the image to gather detailed visual cues necessary for accurate decision-making. In this paper, we propose Zoom Eye, a training-free, model-agnostic tree search algorithm tailored for vision-level reasoning. Zoom Eye treats an image as a hierarchical tree structure, where each child node represents a zoomed-in sub-region of its parent, and the root corresponds to the full image. The algorithm enables MLLMs to simulate human-like zooming behavior by navigating from root to leaf nodes in search of task-relevant visual evidence. We experiment on a series of elaborate high-resolution benchmarks and the results demonstrate that Zoom Eye not only consistently improves the performance of a series of MLLMs with large margin (e.g., InternVL2.5-8B increases by 15.71% and 17.69% on HR-Bench) but also enables small 3-8B MLLMs to outperform strong large models such as GPT-4o.

Yubo Ma, Zhibin Gou, Junheng Hao, Ruochen Xu, Shuohang Wang, Liangming Pan, Yujiu Yang, Yixin Cao, Aixin Sun. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

Large Language Models (LLMs) exhibit emerging in-context learning abilities through prompt engineering. The recent progress in large-scale generative models has further expanded their use in real-world language applications. However, the critical challenge of improving the generalizability and factuality of LLMs in natural language understanding and question answering remains under-explored. While previous in-context learning research has focused on enhancing models to adhere to users' specific instructions and quality expectations, and to avoid undesired outputs, little to no work has explored the use of task-specific fine-tuned Language Models (SLMs) to improve LLMs' in-context learning during the inference stage. Our primary contribution is the establishment of a simple yet effective framework that enhances the reliability of LLMs as it: 1) generalizes out-of-distribution data, 2) elucidates how LLMs benefit from discriminative models, and 3) minimizes hallucinations in generative tasks. Using our proposed plug-in method, enhanced versions of Llama 2 and ChatGPT surpass their original versions regarding generalizability and factuality. We offer a comprehensive suite of resources, including 16 curated datasets, prompts, model checkpoints, and LLM outputs across 9 distinct tasks. Our empirical analysis sheds light on the advantages of incorporating discriminative models into LLMs and highlights the potential of our methodology in fostering more reliable LLMs.

Evaluation of large language models (LLMs) has raised great concerns in the community due to the issue of data contamination. Existing work designed evaluation protocols using well-defined algorithms for specific tasks, which cannot be easily extended to diverse scenarios. Moreover, current evaluation benchmarks can only provide the overall benchmark results and cannot support a fine-grained and multifaceted analysis of LLMs’ abilities. In this paper, we propose meta probing agents (MPA), a general dynamic evaluation protocol inspired by psychometrics to evaluate LLMs. MPA designs the probing and judging agents to automatically transform an original evaluation problem into a new one following psychometric theory on three basic cognitive abilities: language understanding, problem solving, and domain knowledge. These basic abilities are also dynamically configurable, allowing multifaceted analysis. We conducted extensive evaluations using MPA and found that most LLMs achieve poorer performance, indicating room for improvement. Our multifaceted analysis demonstrated the strong correlation between the basic abilities and an implicit Mattew effect on model size, i.e., larger models possess stronger correlations of the abilities. MPA can also be used as a data augmentation approach to enhance LLMs. Code is available at: https://github.com/microsoft/promptbench.

Previous language model pre-training methods have uniformly applied a next-token prediction loss to all training tokens. Challenging this norm, we posit that ''Not all tokens in a corpus are equally important for language model training''. Our initial analysis examines token-level training dynamics of language model, revealing distinct loss patterns for different tokens. Leveraging these insights, we introduce a new language model called Rho-1. Unlike traditional LMs that learn to predict every next token in a corpus, Rho-1 employs Selective Language Modeling (SLM), which selectively trains on useful tokens that aligned with the desired distribution. This approach involves scoring training tokens using a reference model, and then training the language model with a focused loss on tokens with higher scores. When continual continual pretraining on 15B OpenWebMath corpus, Rho-1 yields an absolute improvement in few-shot accuracy of up to 30% in 9 math tasks. After fine-tuning, Rho-1-1B and 7B achieved state-of-the-art results of 40.6% and 51.8% on MATH dataset, respectively - matching DeepSeekMath with only 3% of the pretraining tokens. Furthermore, when continual pretraining on 80B general tokens, Rho-1 achieves 6.8% average enhancement across 15 diverse tasks, increasing both data efficiency and performance of the language model pre-training.

Large language models (LLMs) have achieved unprecedented performances in various applications, yet evaluating them is still challenging. Existing benchmarks are either manually constructed or are automatic, but lack the ability to evaluate the thought process of LLMs with arbitrary complexity. We contend that utilizing existing relational databases based on the entity-relationship (ER) model is a promising approach for constructing benchmarks as they contain structured knowledge that can be used to question LLMs. Unlike knowledge graphs, which are also used to evaluate LLMs, relational databases have integrity constraints that can be used to better construct complex in-depth questions and verify answers: (1) functional dependencies can be used to pinpoint critical keywords that an LLM must know to properly answer a given question containing certain attribute values; and (2) foreign key constraints can be used to join relations and construct multi-hop questions, which can be arbitrarily long and used to debug intermediate answers. We thus propose ERBench, which uses these integrity constraints to convert any database into an LLM benchmark. ERBench supports continuous evaluation as databases change, multimodal questions, and various prompt engineering techniques. In our experiments, we construct LLM benchmarks using databases of multiple domains and make an extensive comparison of contemporary LLMs. We show how ERBench can properly evaluate any LLM by not only checking for answer correctness, but also effectively verifying the rationales by looking for the right keywords.

Yulong Chen, Yang Liu, Ruochen Xu, Ziyi Yang, Chenguang Zhu, Michael Zeng, Yue Zhang. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Pengcheng He, Baolin Peng, Song Wang, Yang Liu, Ruochen Xu, Hany Hassan, Yu Shi, Chenguang Zhu, Wayne Xiong, Michael Zeng, Jianfeng Gao, Xuedong Huang. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

The quality of texts generated by natural language generation (NLG) systems is hard to measure automatically.Conventional referencebased metrics, such as BLEU and ROUGE, have been shown to have relatively low correlation with human judgments, especially for tasks that require creativity and diversity.Recent studies suggest using large language models (LLMs) as reference-free metrics for NLG evaluation, which have the benefit of being applicable to new tasks that lack human references.However, these LLM-based evaluators still have lower human correspondence than medium-size neural evaluators.In this work, we present G-EVAL, a framework of using large language models with chain-of-thoughts (CoT) and a form-filling paradigm, to assess the quality of NLG outputs.We experiment with two generation tasks, text summarization and dialogue generation.We show that G-EVAL with GPT-4 as the backbone model achieves a Spearman correlation of 0.514 with human on summarization task, outperforming all previous methods by a large margin.We also propose analysis on the behavior of LLM-based evaluators, and highlight the potential concern of LLM-based evaluators having a bias towards the LLM-generated texts. 1

Shuohang Wang, Yichong Xu, Yuwei Fang, Yang Liu, Siqi Sun, Ruochen Xu, Chenguang Zhu, Michael Zeng. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.

Vision-language (V+L) pretraining models have achieved great success in supporting multimedia applications by understanding the alignments between images and text. While existing vision-language pretraining models primarily focus on understanding objects in images or entities in text, they often ignore the alignment at the level of events and their argument structures. In this work, we propose a contrastive learning framework to enforce vision-language pretraining models to comprehend events and associated argument (participant) roles. To achieve this, we take advantage of text information extraction technologies to obtain event structural knowledge, and utilize multiple prompt functions to contrast difficult negative descriptions by manipulating event structures. We also design an event graph alignment loss based on optimal transport to capture event argument structures. In addition, we collect a large event-rich dataset (106,875 images) for pretraining, which provides a more challenging image retrieval benchmark to assess the understanding of complicated lengthy sentences11The data and code are publicly available for research purpose in https://github.com/limanling/clip-event.. Experiments show that our zero-shot CLIP-Event outperforms the state-of-the-art supervised model in argument extraction on Multimedia Event Extraction, achieving more than 5% absolute F-score gain in event extraction, as well as significant improvements on a variety of downstream tasks under zero-shot settings.

Paraphrase identification has been formulated as a binary classification task to decide whether two sentences hold a paraphrase relationship.Existing paraphrase datasets only annotate a binary label for each sentence pair.However, after a systematical analysis of existing paraphrase datasets, we found that the degree of paraphrase cannot be well characterized by a single binary label.And the criteria of paraphrase are not even consistent within the same dataset.We hypothesize that such issues would limit the effectiveness of paraphrase models trained on these data.To this end, we propose a novel fine-grained paraphrase annotation schema that labels the minimum spans of tokens in a sentence that don't have the corresponding paraphrases in the other sentence.Under this setting, we frame paraphrasing as a sequence tagging task.We collect 30k sentence pairs in English with the new annotation schema, resulting in the ParaTag dataset.In addition to reporting baseline results on ParaTag using state-of-art language models, we show that ParaTag is especially useful for training an automatic scorer for language generation evaluation.Finally, we train a paraphrase generation model from ParaTag and achieve better data augmentation performance on the GLUE benchmark than other public paraphrasing datasets.1

The goal of this work is to build flexible video-language models that can generalize to various video-to-text tasks from few examples. Existing few-shot video-language learners focus exclusively on the encoder, resulting in the absence of a video-to-text decoder to handle generative tasks. Video captioners have been pretrained on large-scale video-language datasets, but they rely heavily on finetuning and lack the ability to generate text for unseen tasks in a few-shot setting. We propose VidIL, a few-shot Video-language Learner via Image and Language models, which demonstrates strong performance on few-shot video-to-text tasks without the necessity of pretraining or finetuning on any video datasets. We use image-language models to translate the video content into frame captions, object, attribute, and event phrases, and compose them into a temporal-aware template. We then instruct a language model, with a prompt containing a few in-context examples, to generate a target output from the composed content. The flexibility of prompting allows the model to capture any form of text input, such as automatic speech recognition (ASR) transcripts. Our experiments demonstrate the power of language models in understanding videos on a wide variety of video-language tasks, including video captioning, video question answering, video caption retrieval, and video future event prediction. Especially, on video future event prediction, our few-shot model significantly outperforms state-of-the-art supervised models trained on large-scale video datasets.Code and processed data are publicly available for research purposes at https://github.com/MikeWangWZHL/VidIL.

Chenguang Zhu, William Hinthorn, Ruochen Xu, Qingkai Zeng, Michael Zeng, Xuedong Huang, Meng Jiang. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Serious roadside pollution in congested urban areas is an ongoing problem in many densely populated cities. While the current control measures on traffic pollution have reduced particulate matter, new solutions by emerging vehicular technologies can help protect citizens from exhaust-gas emissions. The electric vehicle (EV) is a promising solution to alleviate traffic-induced pollution in urban areas. However, traffic flow will be mixed with fuel vehicles (FVs) and EVs before the FV would be phased out. In this paper, traffic management by road pricing is introduced to reduce the tailpipe emission of the protected area by imposing environmental capacity constraints for network traffic mixed with EVs and FVs. Second-best toll pricing schemes are formulated as side constrained user equilibrium problems for both fixed demand and elastic demand cases. Both the tailpipe emission of FVs and the energy consumption of EVs are assumed to depend nonlinearly on network traffic conditions. Although the EVs do not contribute to the tailpipe emission, all vehicles contribute to congestion externality that induces more emission of FVs. Therefore, both types of vehicles are charged, but the toll on FVs is significantly higher than that on EVs. A new projected dynamics based algorithm is introduced to solve the toll from the Lagrange multiplier associated with the environmental constraint apart from the equilibrium flow. Numerical examples are conducted to evaluate the equilibrium cost and toll, and to analyze the impacts of EV penetration rate on the pricing scheme.

Given the complexity of combinations of tasks, languages, and domains in natural language processing (NLP) research, it is computationally prohibitive to exhaustively test newly proposed models on each possible experimental setting.In this work, we attempt to explore the possibility of gaining plausible judgments of how well an NLP model can perform under an experimental setting, without actually training or testing the model.To do so, we build regression models to predict the evaluation score of an NLP experiment given the experimental settings as input.Experimenting on 9 different NLP tasks, we find that our predictors can produce meaningful predictions over unseen languages and different modeling architectures, outperforming reasonable baselines as well as human experts.Going further, we outline how our predictor can be used to find a small subset of representative experiments that should be run in order to obtain plausible predictions for all other experimental settings.1

The use of machine learning for NLP generally requires resources for training. Tasks performed in a low-resource language usually rely on labeled data in another, typically resource-rich, language. However, there might not be enough labeled data even in a resource-rich language such as English. In such cases, one approach is to use a hand-crafted approach that utilizes only a small bilingual dictionary with minimal manual verification to create distantly supervised data. Another is to explore typical machine learning techniques, for example adversarial training of bilingual word representations. We find that in event-type detection task—the task to classify [parts of] documents into a fixed set of labels—they give about the same performance. We explore ways in which the two methods can be complementary and also see how to best utilize a limited budget for manual annotation to maximize performance gain.

Cross-lingual transfer of word embeddings aims to establish the semantic mappings among words in different languages by learning the transformation functions over the corresponding word embedding spaces.Successfully solving this problem would benefit many downstream tasks such as to translate text classification models from resource-rich languages (e.g.English) to low-resource languages.Supervised methods for this problem rely on the availability of cross-lingual supervision, either using parallel corpora or bilingual lexicons as the labeled data for training, which may not be available for many low resource languages.This paper proposes an unsupervised learning approach that does not require any cross-lingual labeled data.Given two monolingual word embedding spaces for any language pair, our algorithm optimizes the transformation functions in both directions simultaneously based on distributional matching as well as minimizing the backtranslation losses.We use a neural network implementation to calculate the Sinkhorn distance, a well-defined distributional similarity measure, and optimize our objective through back-propagation.Our evaluation on benchmark datasets for bilingual lexicon induction and cross-lingual word similarity prediction shows stronger or competitive performance of the proposed method compared to other stateof-the-art supervised and unsupervised baseline methods over many language pairs.

Cross-lingual text classification(CLTC) is the task of classifying documents written in different languages into the same taxonomy of categories.This paper presents a novel approach to CLTC that builds on model distillation, which adapts and extends a framework originally proposed for model compression.Using soft probabilistic predictions for the documents in a label-rich language as the (induced) supervisory labels in a parallel corpus of documents, we train classifiers successfully for new languages in which labeled training data are not available.An adversarial feature adaptation technique is also applied during the model training to reduce distribution mismatch.We conducted experiments on two benchmark CLTC datasets, treating English as the source language and German, French, Japan and Chinese as the unlabeled target languages.The proposed approach had the advantageous or comparable performance of the other state-of-art methods.

Software architects inhabit a complex, rapidly evolving technological landscape. An ever growing collection of competing architecturally significant technologies, ranging from distributed databases to middleware and cloud platforms, makes rigorously comparing alternatives and selecting appropriate solutions a daunting engineering task. To address this problem, we envisage an ecosystem of curated, automatically updated knowledge bases that enable straightforward and streamlined technical comparisons of related products. These knowledge bases would emulate engineering handbooks that are commonly found in other engineering disciplines. As a first step towards this vision, we have built a curated knowledge base for comparing distributed databases based on a semantically defined feature taxonomy. We report in this paper on the initial results of using supervised machine learning to assist with knowledge base curation. Our results show immense promise in recommending Web pages that are highly relevant to curators. We also describe the major obstacles, both practical and scientific, that our work has uncovered. These must be overcome by future research in order to make our vision of curated knowledge bases a reality.

Cross-lingual text classification (CLTC) refers to the task of classifying documents in different languages into the same taxonomy of categories. An open challenge in CLTC is to classify documents for the languages where labeled training data are not available. Existing approaches rely on the availability of either high-quality machine translation of documents (to the languages where massively training data are available), or rich bilingual dictionaries for effective translation of trained classification models (to the languages where labeled training data are lacking). This paper studies the CLTC challenge under the assumption that neither condition is met. That is, we focus on the problem of translating classification models with highly incomplete bilingual dictionaries. Specifically, we propose two new approaches that combines unsupervised word embedding in different languages, supervised mapping of embedded words across languages, and probabilistic translation of classification models. The approaches show significant performance improvement in CLTC on a benchmark corpus of Reuters news stories (RCV1/RCV2) in English, Spanish, German, French and Chinese and an internal dataset in Uzbek, compared to representative baseline methods using conventional bilingual dictionaries or highly incomplete ones.

Event extraction has become one of the most important topics in information extraction, but to date, there is very limited work on leveraging cross-lingual training to boost performance. We propose a new event extraction approach that trains on multiple languages using a combination of both language-dependent and language-independent features, with particular focus on the case where target domain training data is of very limited size. We show empirically that multilingual training can boost performance for the tasks of event trigger extraction and event argument extraction on the Chinese ACE 2005 dataset.