Jesse Atuhurra

Jesse Atuhurra, Iqra Ali, Tomoya Iwakura et al.

Vision Language Models (VLMs) are pivotal for advancing perception in intelligent agents. Yet, evaluation of VLMs remains limited to predominantly English-centric benchmarks in which the image-text pairs comprise short texts. To evaluate VLM fine-grained abilities, in four languages under long-text settings, we introduce a novel multilingual benchmark VLURes featuring eight vision-and-language tasks, and a pioneering unrelatedness task, to probe the fine-grained Visual and Linguistic Understanding capabilities of VLMs across English, Japanese, and low-resource languages, Swahili, and Urdu. Our datasets, curated from web resources in the target language, encompass ten diverse image categories and rich textual context, introducing valuable vision-language resources for Swahili and Urdu. By prompting VLMs to generate responses and rationales, evaluated automatically and by native speakers, we uncover performance disparities across languages and tasks critical to intelligent agents, such as object recognition, scene understanding, and relationship understanding. We conducted evaluations of ten VLMs with VLURes. The best performing model, GPT-4o, achieves an overall accuracy of 90.8% and lags human performance by 6.7%, though the gap is larger for open-source models. The gap highlights VLURes' critical role in developing intelligent agents to tackle multi-modal visual reasoning.

Jesse Atuhurra, Hidetaka Kamigaito, Taro Watanabe et al.

We introduce J-ORA, a novel multimodal dataset that bridges the gap in robot perception by providing detailed object attribute annotations within Japanese human-robot dialogue scenarios. J-ORA is designed to support three critical perception tasks, object identification, reference resolution, and next-action prediction, by leveraging a comprehensive template of attributes (e.g., category, color, shape, size, material, and spatial relations). Extensive evaluations with both proprietary and open-source Vision Language Models (VLMs) reveal that incorporating detailed object attributes substantially improves multimodal perception performance compared to without object attributes. Despite the improvement, we find that there still exists a gap between proprietary and open-source VLMs. In addition, our analysis of object affordances demonstrates varying abilities in understanding object functionality and contextual relationships across different VLMs. These findings underscore the importance of rich, context-sensitive attribute annotations in advancing robot perception in dynamic environments. See project page at https://jatuhurrra.github.io/J-ORA/.

Jesse Atuhurra, Takanori Hara, Yuanyu Zhang et al.

With the rapidly spreading usage of Internet of Things (IoT) devices, a network intrusion detection system (NIDS) plays an important role in detecting and protecting various types of attacks in the IoT network. To evaluate the robustness of the NIDS in the IoT network, the existing work proposed a realistic botnet dataset in the IoT network (Bot-IoT dataset) and applied it to machine learning-based anomaly detection. This dataset contains imbalanced normal and attack packets because the number of normal packets is much smaller than that of attack ones. The nature of imbalanced data may make it difficult to identify the minority class correctly. In this thesis, to address the class imbalance problem in the Bot-IoT dataset, we propose a binary classification method with synthetic minority over-sampling techniques (SMOTE). The proposed classifier aims to detect attack packets and overcome the class imbalance problem using the SMOTE algorithm. Through numerical results, we demonstrate the proposed classifier's fundamental characteristics and the impact of imbalanced data on its performance.

Jesse Atuhurra, Hiroyuki Shindo, Hidetaka Kamigaito et al.

Many attempts have been made in multilingual NLP to ensure that pre-trained language models, such as mBERT or GPT2 get better and become applicable to low-resource languages. To achieve multilingualism for pre-trained language models (PLMs), we need techniques to create word embeddings that capture the linguistic characteristics of any language. Tokenization is one such technique because it allows for the words to be split based on characters or subwords, creating word embeddings that best represent the structure of the language. Creating such word embeddings is essential to applying PLMs to other languages where the model was not trained, enabling multilingual NLP. However, most PLMs use generic tokenization methods like BPE, wordpiece, or unigram which may not suit specific languages. We hypothesize that tokenization based on syllables within the input text, which we call syllable tokenization, should facilitate the development of syllable-aware language models. The syllable-aware language models make it possible to apply PLMs to languages that are rich in syllables, for instance, Swahili. Previous works introduced subword tokenization. Our work extends such efforts. Notably, we propose a syllable tokenizer and adopt an experiment-centric approach to validate the proposed tokenizer based on the Swahili language. We conducted text-generation experiments with GPT2 to evaluate the effectiveness of the syllable tokenizer. Our results show that the proposed syllable tokenizer generates syllable embeddings that effectively represent the Swahili language.

Jesse Atuhurra

The emergence of large language models (LLM) and, consequently, vision language models (VLM) has ignited new imaginations among robotics researchers. At this point, the range of applications to which LLM and VLM can be applied in human-robot interaction (HRI), particularly socially assistive robots (SARs), is unchartered territory. However, LLM and VLM present unprecedented opportunities and challenges for SAR integration. We aim to illuminate the opportunities and challenges when roboticists deploy LLM and VLM in SARs. First, we conducted a meta-study of more than 250 papers exploring 1) major robots in HRI research and 2) significant applications of SARs, emphasizing education, healthcare, and entertainment while addressing 3) societal norms and issues like trust, bias, and ethics that the robot developers must address. Then, we identified 4) critical components of a robot that LLM or VLM can replace while addressing the 5) benefits of integrating LLM into robot designs and the 6) risks involved. Finally, we outline a pathway for the responsible and effective adoption of LLM or VLM into SARs, and we close our discussion by offering caution regarding this deployment.

Jesse Atuhurra, Seiveright Cargill Dujohn, Hidetaka Kamigaito et al.

Natural language processing (NLP) practitioners are leveraging large language models (LLM) to create structured datasets from semi-structured and unstructured data sources such as patents, papers, and theses, without having domain-specific knowledge. At the same time, ecological experts are searching for a variety of means to preserve biodiversity. To contribute to these efforts, we focused on endangered species and through in-context learning, we distilled knowledge from GPT-4. In effect, we created datasets for both named entity recognition (NER) and relation extraction (RE) via a two-stage process: 1) we generated synthetic data from GPT-4 of four classes of endangered species, 2) humans verified the factual accuracy of the synthetic data, resulting in gold data. Eventually, our novel dataset contains a total of 3.6K sentences, evenly divided between 1.8K NER and 1.8K RE sentences. The constructed dataset was then used to fine-tune both general BERT and domain-specific BERT variants, completing the knowledge distillation process from GPT-4 to BERT, because GPT-4 is resource intensive. Experiments show that our knowledge transfer approach is effective at creating a NER model suitable for detecting endangered species from texts.

Jesse Atuhurra, Hidetaka Kamigaito, Taro Watanabe et al.

Dialogue agents, which perform specific tasks, are part of the long-term goal of NLP researchers to build intelligent agents that communicate with humans in natural language. Such systems should adapt easily from one domain to another to assist users in completing tasks. Researchers have developed a broad range of techniques, objectives, and datasets for intent classification to achieve such systems. Despite the progress in developing intent classification systems (ICS), a systematic review of the progress from a technical perspective is yet to be conducted. In effect, important implementation details of intent classification remain restricted and unclear, making it hard for natural language processing (NLP) researchers to develop new methods. To fill this gap, we review contemporary works in intent classification. Specifically, we conduct a thorough technical review of the datasets, domains, tasks, and methods needed to train the intent classification part of dialogue systems. Our structured analysis describes why intent classification is difficult and studies the limitations to domain adaptation while presenting opportunities for future work.

Jesse Atuhurra, Iqra Ali, Tatsuya Hiraoka et al.

Large language models (LLMs) have increased interest in vision language models (VLMs), which process image-text pairs as input. Studies investigating the visual understanding ability of VLMs have been proposed, but such studies are still preliminary because existing datasets do not permit a comprehensive evaluation of the fine-grained visual linguistic abilities of VLMs across multiple languages. To further explore the strengths of VLMs, such as GPT-4V \cite{openai2023GPT4}, we developed new datasets for the systematic and qualitative analysis of VLMs. Our contribution is four-fold: 1) we introduced nine vision-and-language (VL) tasks (including object recognition, image-text matching, and more) and constructed multilingual visual-text datasets in four languages: English, Japanese, Swahili, and Urdu through utilizing templates containing \textit{questions} and prompting GPT4-V to generate the \textit{answers} and the \textit{rationales}, 2) introduced a new VL task named \textit{unrelatedness}, 3) introduced rationales to enable human understanding of the VLM reasoning process, and 4) employed human evaluation to measure the suitability of proposed datasets for VL tasks. We show that VLMs can be fine-tuned on our datasets. Our work is the first to conduct such analyses in Swahili and Urdu. Also, it introduces \textit{rationales} in VL analysis, which played a vital role in the evaluation.

Jesse Atuhurra, Hidetaka Kamigaito, Hiroki Ouchi et al.

Adapting named entity recognition (NER) methods to new domains poses significant challenges. We introduce RapidNER, a framework designed for the rapid deployment of NER systems through efficient dataset construction. RapidNER operates through three key steps: (1) extracting domain-specific sub-graphs and triples from a general knowledge graph, (2) collecting and leveraging texts from various sources to build the NERsocial dataset, which focuses on entities typical in human-robot interaction, and (3) implementing an annotation scheme using Elasticsearch (ES) to enhance efficiency. NERsocial, validated by human annotators, includes six entity types, 153K tokens, and 99.4K sentences, demonstrating RapidNER's capability to expedite dataset creation.

Jesse Atuhurra, Hidetaka Kamigaito

Natural language processing (NLP) has grown significantly since the advent of the Transformer architecture. Transformers have given birth to pre-trained large language models (PLMs). There has been tremendous improvement in the performance of NLP systems across several tasks. NLP systems are on par or, in some cases, better than humans at accomplishing specific tasks. However, it remains the norm that \emph{better quality datasets at the time of pretraining enable PLMs to achieve better performance, regardless of the task.} The need to have quality datasets has prompted NLP researchers to continue creating new datasets to satisfy particular needs. For example, the two top NLP conferences, ACL and EMNLP, accepted ninety-two papers in 2022, introducing new datasets. This work aims to uncover the trends and insights mined within these datasets. Moreover, we provide valuable suggestions to researchers interested in curating datasets in the future.