Considerable efforts have been invested in augmenting the role-playing proficiency of open-source large language models (LLMs) by emulating proprietary counterparts. Nevertheless, we posit that LLMs inherently harbor role-play capabilities, owing to the extensive knowledge of characters and potential dialogues ingrained in their vast training corpora. Thus, in this study, we introduce Ditto, a self-alignment method for role-play. Ditto capitalizes on character knowledge, encouraging an instruction-following LLM to simulate role-play dialogues as a variant of reading comprehension. This method creates a role-play training set comprising 4,000 characters, surpassing the scale of currently available datasets by tenfold regarding the number of roles. Subsequently, we fine-tune the LLM using this self-generated dataset to augment its role-playing capabilities. Upon evaluating our meticulously constructed and reproducible role-play benchmark and the roleplay subset of MT-Bench, Ditto, in various parameter scales, consistently maintains a consistent role identity and provides accurate role-specific knowledge in multi-turn role-play conversations. Notably, it outperforms all open-source role-play baselines, showcasing performance levels comparable to advanced proprietary chatbots. Furthermore, we present the first comprehensive cross-supervision alignment experiment in the role-play domain, revealing that the intrinsic capabilities of LLMs confine the knowledge within role-play. Meanwhile, the role-play styles can be easily acquired with the guidance of smaller models. We open-source related resources at https://github.com/OFA-Sys/Ditto.
Vision-language models (VLMs) excel in zero-shot recognition but exhibit drastically imbalanced performance across visual concepts. For example, CLIP, despite an impressive mean zero-shot accuracy on ImageNet (72.7%), yields $<$10% on ten concepts (e.g., gyromitra and night snake), presumably, because these concepts are under-represented in VLMs' imbalanced pretraining data. Yet, assessing this imbalance is challenging as it is non-trivial to calculate the frequency of specific concepts within VLMs' large-scale pretraining data. Our work makes the first attempt to measure the concept frequency by analyzing pretraining texts. We use off-the-shelf language models to help count relevant texts that contain synonyms of the given concepts and resolve linguistic ambiguity. We confirm that popular VLM datasets like LAION indeed exhibit long-tailed concept distributions, which strongly correlate with per-class accuracies. Further, contemporary multimodal systems, e.g., visual chatbots and text-to-image generators, also struggle with the rare concepts identified by our method. To mitigate VLMs' imbalanced performance in zero-shot recognition, we propose REtrieval-Augmented Learning REAL. First, instead of prompting VLMs using the original class names, REAL uses their most frequent synonyms found in VLMs' pretraining texts. This already outperforms human-engineered and LLM-generated prompts over nine benchmark datasets, likely because VLMs have seen more images associated with the frequently used synonyms. Second, REAL uses all the concept synonyms to retrieve a small, class-balanced set of pretraining data to train a robust classifier. REAL surpasses the recent retrieval-augmented solution REACT, using 400x less storage and 10,000x less training time!
Chatbots, the common moniker for collaborative assistants, are Artificial Intelligence (AI) software that enables people to naturally interact with them to get tasks done. Although chatbots have been studied since the dawn of AI, they have particularly caught the imagination of the public and businesses since the launch of easy-to-use and general-purpose Large Language Model-based chatbots like ChatGPT. As businesses look towards chatbots as a potential technology to engage users, who may be end customers, suppliers, or even their own employees, proper testing of chatbots is important to address and mitigate issues of trust related to service or product performance, user satisfaction and long-term unintended consequences for society. This paper reviews current practices for chatbot testing, identifies gaps as open problems in pursuit of user trust, and outlines a path forward.
As artificial intelligence (AI) becomes more widespread, one question that arises is how human-AI interaction might impact human-human interaction. Chatbots, for example, are increasingly used as social companions, but little is known about how their use impacts human relationships. A common hypothesis is that these companion bots are detrimental to social health by harming or replacing human interaction. To understand how companion bots impact social health, we studied people who used companion bots and people who did not. Contrary to expectations, companion bot users indicated that these relationships were beneficial to their social health, whereas nonusers viewed them as harmful. Another common assumption is that people perceive conscious, humanlike AI as disturbing and threatening. Among both users and nonusers, however, we found the opposite: perceiving companion bots as more conscious and humanlike correlated with more positive opinions and better social health benefits. Humanlike bots may aid social health by supplying reliable and safe interactions, without necessarily harming human relationships.
Chatbots have become one of the main pathways for the delivery of business automation tools. Multi-agent systems offer a framework for designing chatbots at scale, making it easier to support complex conversations that span across multiple domains as well as enabling developers to maintain and expand their capabilities incrementally over time. However, multi-agent systems complicate the natural language understanding (NLU) of user intents, especially when they rely on decentralized NLU models: some utterances (termed single intent) may invoke a single agent while others (termed multi-intent) may explicitly invoke multiple agents. Without correctly parsing multi-intent inputs, decentralized NLU approaches will not achieve high prediction accuracy. In this paper, we propose an efficient parsing and orchestration pipeline algorithm to service multi-intent utterances from the user in the context of a multi-agent system. Our proposed approach achieved comparable performance to competitive deep learning models on three different datasets while being up to 48 times faster.
With the rise of human-machine communication, machines are increasingly designed with humanlike characteristics, such as gender, which can inadvertently trigger cognitive biases. Many conversational agents (CAs), such as voice assistants and chatbots, default to female personas, leading to concerns about perpetuating gender stereotypes and inequality. Critiques have emerged regarding the potential objectification of females and reinforcement of gender stereotypes by these technologies. This research, situated in conversational AI design, aims to delve deeper into the impacts of gender biases in human-CA interactions. From a behavioral and communication research standpoint, this program focuses not only on perceptions but also the linguistic styles of users when interacting with CAs, as previous research has rarely explored. It aims to understand how pre-existing gender biases might be triggered by CAs' gender designs. It further investigates how CAs' gender designs may reinforce gender biases and extend them to human-human communication. The findings aim to inform ethical design of conversational agents, addressing whether gender assignment in CAs is appropriate and how to promote gender equality in design.
Autonomous conversational agents, i.e. chatbots, are becoming an increasingly common mechanism for enterprises to provide support to customers and partners. In order to rate chatbots, especially ones powered by Generative AI tools like Large Language Models (LLMs) we need to be able to accurately assess their performance. This is where chatbot benchmarking becomes important. In this paper, we propose the use of a novel benchmark that we call the E2E (End to End) benchmark, and show how the E2E benchmark can be used to evaluate accuracy and usefulness of the answers provided by chatbots, especially ones powered by LLMs. We evaluate an example chatbot at different levels of sophistication based on both our E2E benchmark, as well as other available metrics commonly used in the state of art, and observe that the proposed benchmark show better results compared to others. In addition, while some metrics proved to be unpredictable, the metric associated with the E2E benchmark, which uses cosine similarity performed well in evaluating chatbots. The performance of our best models shows that there are several benefits of using the cosine similarity score as a metric in the E2E benchmark.
Artificial Intelligence (AI) for supporting healthcare services has never been more necessitated than by the recent global pandemic. Here, we review the state-of-the-art in AI-enabled Chatbots in healthcare proposed during the last 10 years (2013-2023). The focus on AI-enabled technology is because of its potential for enhancing the quality of human-machine interaction via Chatbots, reducing dependence on human-human interaction and saving man-hours. Our review indicates that there are a handful of (commercial) Chatbots that are being used for patient support, while there are others (non-commercial) that are in the clinical trial phases. However, there is a lack of trust on this technology regarding patient safety and data protection, as well as a lack of wider awareness on its benefits among the healthcare workers and professionals. Also, patients have expressed dissatisfaction with Natural Language Processing (NLP) skills of the Chatbots in comparison to humans. Notwithstanding the recent introduction of ChatGPT that has raised the bar for the NLP technology, this Chatbot cannot be trusted with patient safety and medical ethics without thorough and rigorous checks to serve in the `narrow' domain of assistive healthcare. Our review suggests that to enable deployment and integration of AI-enabled Chatbots in public health services, the need of the hour is: to build technology that is simple and safe to use; to build confidence on the technology among: (a) the medical community by focussed training and development; (b) the patients and wider community through outreach.
Personalizing conversational agents can enhance the quality of conversations and increase user engagement. However, they often lack external knowledge to appropriately tend to a user's persona. This is particularly crucial for practical applications like mental health support, nutrition planning, culturally sensitive conversations, or reducing toxic behavior in conversational agents. To enhance the relevance and comprehensiveness of personalized responses, we propose using a two-step approach that involves (1) selectively integrating user personas and (2) contextualizing the response with supplementing information from a background knowledge source. We develop K-PERM (Knowledge-guided PErsonalization with Reward Modulation), a dynamic conversational agent that combines these elements. K-PERM achieves state-of-the-art performance on the popular FoCus dataset, containing real-world personalized conversations concerning global landmarks. We show that using responses from K-PERM can improve performance in state-of-the-art LLMs (GPT 3.5) by 10.5%, highlighting the impact of K-PERM for personalizing chatbots.
We address the integration of storytelling and Large Language Models (LLMs) to develop engaging and believable Social Chatbots (SCs) in community settings. Motivated by the potential of fictional characters to enhance social interactions, we introduce Storytelling Social Chatbots (SSCs) and the concept of story engineering to transform fictional game characters into "live" social entities within player communities. Our story engineering process includes three steps: (1) Character and story creation, defining the SC's personality and worldview, (2) Presenting Live Stories to the Community, allowing the SC to recount challenges and seek suggestions, and (3) Communication with community members, enabling interaction between the SC and users. We employed the LLM GPT-3 to drive our SSC prototypes, "David" and "Catherine," and evaluated their performance in an online gaming community, "DE (Alias)," on Discord. Our mixed-method analysis, based on questionnaires (N=15) and interviews (N=8) with community members, reveals that storytelling significantly enhances the engagement and believability of SCs in community settings.