PyTorchKR Papers

최신 논문

총 758편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2401.07810

Consolidating Strategies for Countering Hate Speech Using Persuasive Dialogues

Hateful comments are prevalent on social media platforms. Although tools for automatically detecting, flagging, and blocking such false, offensive, and harmful content online have lately matured, such reactive and brute force methods alone provide short-term and superficial remedies while the perpetrators persist. With the public availability of large language models which can generate articulate synthetic and engaging content at scale, there are concerns about the rapid growth of dissemination of such malicious content on the web. There is now a need to focus on deeper, long-term solutions that involve engaging with the human perpetrator behind the source of the content to change their viewpoint or at least bring down the rhetoric using persuasive means. To do that, we propose defining and experimenting with controllable strategies for generating counter-arguments to hateful comments in online conversations. We experiment with controlling response generation using features based on (i) argument structure and reasoning-based Walton argument schemes, (ii) counter-argument speech acts, and (iii) human characteristics-based qualities such as Big-5 personality traits and human values. Using automatic and human evaluations, we determine the best combination of features that generate fluent, argumentative, and logically sound arguments for countering hate. We further share the developed computational models for automatically annotating text with such features, and a silver-standard annotated version of an existing hate speech dialog corpora.

arXiv:2401.06541

Medical Dialogue Generation via Intuitive-then-Analytical Differential Diagnosis

Medical dialogue systems have attracted growing research attention as they have the potential to provide rapid diagnoses, treatment plans, and health consultations. In medical dialogues, a proper diagnosis is crucial as it establishes the foundation for future consultations. Clinicians typically employ both intuitive and analytic reasoning to formulate a differential diagnosis. This reasoning process hypothesizes and verifies a variety of possible diseases and strives to generate a comprehensive and rigorous diagnosis. However, recent studies on medical dialogue generation have overlooked the significance of modeling a differential diagnosis, which hinders the practical application of these systems. To address the above issue, we propose a medical dialogue generation framework with the Intuitive-then-Analytic Differential Diagnosis (IADDx). Our method starts with a differential diagnosis via retrieval-based intuitive association and subsequently refines it through a graph-enhanced analytic procedure. The resulting differential diagnosis is then used to retrieve medical knowledge and guide response generation. Experimental results on two datasets validate the efficacy of our method. Besides, we demonstrate how our framework assists both clinicians and patients in understanding the diagnostic process, for instance, by producing intermediate results and graph-based diagnosis paths.

arXiv:2401.02297

Are LLMs Robust for Spoken Dialogues?

Large Pre-Trained Language Models have demonstrated state-of-the-art performance in different downstream tasks, including dialogue state tracking and end-to-end response generation. Nevertheless, most of the publicly available datasets and benchmarks on task-oriented dialogues focus on written conversations. Consequently, the robustness of the developed models to spoken interactions is unknown. In this work, we have evaluated the performance of LLMs for spoken task-oriented dialogues on the DSTC11 test sets. Due to the lack of proper spoken dialogue datasets, we have automatically transcribed a development set of spoken dialogues with a state-of-the-art ASR engine. We have characterized the ASR-error types and their distributions and simulated these errors in a large dataset of dialogues. We report the intrinsic (perplexity) and extrinsic (human evaluation) performance of fine-tuned GPT-2 and T5 models in two subtasks of response generation and dialogue state tracking, respectively. The results show that LLMs are not robust to spoken noise by default, however, fine-tuning/training such models on a proper dataset of spoken TODs can result in a more robust performance.

arXiv:2312.17748공식 코드 · Papers with Code

K-PERM: Personalized Response Generation Using Dynamic Knowledge Retrieval and Persona-Adaptive Queries

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.

arXiv:2312.16864

OmniDialog: An Omnipotent Pre-training Model for Task-Oriented Dialogue System

Pre-trained conversation models (PCMs) have demonstrated remarkable results in task-oriented dialogue (TOD) systems. Many PCMs focus predominantly on dialogue management tasks like dialogue state tracking, dialogue generation tasks like response generation, or both. However, the existing PCMs seldom consider dialogue comprehension tasks, such as dialogue question answering and summarization tasks. These tasks allow PCMs to glean dialogue context from various angles. This observation naturally raises the question: Can the performance of downstream dialogue tasks be enhanced if a PCM is pre-trained on dialogue management, generation, and comprehension tasks? To investigate this, we proposed an Omnipotent Dialogue pre-training model (OmniDialog). It unifies these three dialogue tasks into a monolithic framework by multi-task learning, fostering inter-task communication. The pre-training corpus of OmniDialog spans $\mathbf{7}$ dialogue-focused tasks, drawing from $\mathbf{15}$ datasets and encompassing over $\mathbf{3.2}$ million dialogue utterances. To our knowledge, OmniDialog is a pioneering PCM pre-trained across dialogue management, generation, and comprehension domains. We evaluated its performance across four tasks: dialogue summarization, end-to-end dialogue modeling, dialogue state tracking, and intent classification. The results underscore its efficacy in domain transfer learning, low-resource, and full-dataset scenarios. Furthermore, to glean a nuanced understanding of OmniDialog's strengths and potential pitfalls, we designed a fine-grained analysis framework for dialogue-centric tasks. Experimental results show that the OmniDialog is good at hard samples, such as long dialogues and lengthy responses.

arXiv:2312.15198

Do LLM Agents Exhibit Social Behavior?

As LLMs increasingly take on roles in human-AI interactions and autonomous AI systems, understanding their social behavior becomes important for informed use and continuous improvement. However, their behaviors in social interactions with humans and other agents, as well as the mechanisms shaping their responses, remain underexplored. To address this gap, we introduce a novel probabilistic framework, State-Understanding-Value-Action (SUVA), to systematically analyze LLM responses in social contexts based on their textual outputs (i.e., utterances). Using canonical behavioral economics games and social preference concepts relatable to LLM users, SUVA assesses LLMs' social behavior through both their final decisions and the response generation processes leading to those decisions. Our analysis of eight LLMs -- including two GPT, four LLaMA, and two Mistral models -- suggests that most models do not generate decisions aligned solely with self-interest; instead, they often produce responses that reflect social welfare considerations and display patterns consistent with direct and indirect reciprocity. Additionally, higher-capacity models more frequently display group identity effects. The SUVA framework also provides explainable tools -- including tree-based visualizations and probabilistic dependency analysis -- to elucidate how factors in LLMs' utterance-based reasoning influence their decisions. We demonstrate that utterance-based reasoning reliably predicts LLMs' final actions; references to altruism, fairness, and cooperation in the reasoning increase the likelihood of prosocial actions, while mentions of self-interest and competition reduce them. Overall, our framework enables practitioners to assess LLMs for applications involving social interactions, and provides researchers with a structured method to interpret how LLM behavior arises from utterance-based reasoning.

arXiv:2312.15006

Assessing the Impact of Prompting Methods on ChatGPT's Mathematical Capabilities

This study critically evaluates the efficacy of prompting methods in enhancing the mathematical reasoning capability of large language models (LLMs). The investigation uses three prescriptive prompting methods - simple, persona, and conversational prompting - known for their effectiveness in enhancing the linguistic tasks of LLMs. We conduct this analysis on OpenAI's LLM chatbot, ChatGPT-3.5, on extensive problem sets from the MATH, GSM8K, and MMLU datasets, encompassing a broad spectrum of mathematical challenges. A grading script adapted to each dataset is used to determine the effectiveness of these prompting interventions in enhancing the model's mathematical analysis power. Contrary to expectations, our empirical analysis reveals that none of the investigated methods consistently improves over ChatGPT-3.5's baseline performance, with some causing significant degradation. Our findings suggest that prompting strategies do not necessarily generalize to new domains, in this study failing to enhance mathematical performance.

arXiv:2312.10371

K-ESConv: Knowledge Injection for Emotional Support Dialogue Systems via Prompt Learning

Automatic psychological counseling requires mass of professional knowledge that can be found in online counseling forums. Motivated by this, we propose K-ESConv, a novel prompt learning based knowledge injection method for emotional support dialogue system, transferring forum knowledge to response generation. We evaluate our model on an emotional support dataset ESConv, where the model retrieves and incorporates knowledge from external professional emotional Q\&A forum. Experiment results show that the proposed method outperforms existing baselines on both automatic evaluation and human evaluation, which shows that our approach significantly improves the correlation and diversity of responses and provides more comfort and better suggestion for the seeker.

arXiv:2312.10355

CoAScore: Chain-of-Aspects Prompting for NLG Evaluation

Recently, natural language generation (NLG) evaluation has shifted from a single-aspect to a multi-aspect paradigm, allowing for a more accurate assessment. Large language models (LLMs) achieve superior performance on various NLG evaluation tasks. However, current work often employs the LLM to independently evaluate different aspects, which largely ignores the rich correlation between various aspects. To fill this research gap, in this work, we propose an NLG evaluation metric called CoAScore. Powered by LLMs, the CoAScore utilizes multi-aspect knowledge through a CoA (\textbf{C}hain-\textbf{o}f-\textbf{A}spects) prompting framework when assessing the quality of a certain aspect. Specifically, for a given aspect to evaluate, we first prompt the LLM to generate a chain of aspects that are relevant to the target aspect and could be useful for the evaluation. We then collect evaluation scores for each generated aspect, and finally, leverage the knowledge of these aspects to improve the evaluation of the target aspect. We evaluate CoAScore across five NLG evaluation tasks (e.g., summarization, dialog response generation, etc) and nine aspects (e.g., overall quality, relevance, coherence, etc). Our experimental findings highlight that, in comparison to individual aspect evaluation, CoAScore exhibits a higher correlation with human judgments. This improvement significantly outperforms existing unsupervised evaluation metrics, whether for assessing overall quality or other aspects. We also conducted extensive ablation studies to validate the effectiveness of the three stages within the CoAScore framework and conducted case studies to show how the LLM performs in these stages. Our code and scripts are available.

arXiv:2312.10967공식 코드 · Papers with Code

Knowledge Graphs and Pre-trained Language Models enhanced Representation Learning for Conversational Recommender Systems

Conversational recommender systems (CRS) utilize natural language interactions and dialogue history to infer user preferences and provide accurate recommendations. Due to the limited conversation context and background knowledge, existing CRSs rely on external sources such as knowledge graphs to enrich the context and model entities based on their inter-relations. However, these methods ignore the rich intrinsic information within entities. To address this, we introduce the Knowledge-Enhanced Entity Representation Learning (KERL) framework, which leverages both the knowledge graph and a pre-trained language model to improve the semantic understanding of entities for CRS. In our KERL framework, entity textual descriptions are encoded via a pre-trained language model, while a knowledge graph helps reinforce the representation of these entities. We also employ positional encoding to effectively capture the temporal information of entities in a conversation. The enhanced entity representation is then used to develop a recommender component that fuses both entity and contextual representations for more informed recommendations, as well as a dialogue component that generates informative entity-related information in the response text. A high-quality knowledge graph with aligned entity descriptions is constructed to facilitate our study, namely the Wiki Movie Knowledge Graph (WikiMKG). The experimental results show that KERL achieves state-of-the-art results in both recommendation and response generation tasks.

arXiv:2312.08702

Rational Sensibility: LLM Enhanced Empathetic Response Generation Guided by Self-presentation Theory

The development of Large Language Models (LLMs) provides human-centered Artificial General Intelligence (AGI) with a glimmer of hope. Empathy serves as a key emotional attribute of humanity, playing an irreplaceable role in human-centered AGI. Despite numerous researches aim to improve the cognitive empathy of models by incorporating external knowledge, there has been limited attention on the sensibility and rationality of the conversation itself, which are vital components of the empathy. However, the rationality information within the conversation is restricted, and previous methods of extending knowledge are subject to semantic conflict and single-role view. In this paper, we design an innovative encoder module inspired by self-presentation theory in sociology, which specifically processes sensibility and rationality sentences in dialogues. And we employ a LLM as a rational brain to decipher profound logical information preserved within the conversation, which assists our model in assessing the balance between sensibility and rationality to produce high-quality empathetic response. Experimental results demonstrate that our model outperforms other methods in both automatic and human evaluations.

arXiv:2312.04668공식 코드 · Papers with Code

TOD-Flow: Modeling the Structure of Task-Oriented Dialogues

Task-Oriented Dialogue (TOD) systems have become crucial components in interactive artificial intelligence applications. While recent advances have capitalized on pre-trained language models (PLMs), they exhibit limitations regarding transparency and controllability. To address these challenges, we propose a novel approach focusing on inferring the TOD-Flow graph from dialogue data annotated with dialog acts, uncovering the underlying task structure in the form of a graph. The inferred TOD-Flow graph can be easily integrated with any dialogue model to improve its prediction performance, transparency, and controllability. Our TOD-Flow graph learns what a model can, should, and should not predict, effectively reducing the search space and providing a rationale for the model's prediction. We show that the proposed TOD-Flow graph better resembles human-annotated graphs compared to prior approaches. Furthermore, when combined with several dialogue policies and end-to-end dialogue models, we demonstrate that our approach significantly improves dialog act classification and end-to-end response generation performance in the MultiWOZ and SGD benchmarks. Code available at: https://github.com/srsohn/TOD-Flow

arXiv:2312.01886공식 코드 · Papers with Code

InstructTA: Instruction-Tuned Targeted Attack for Large Vision-Language Models

Large vision-language models (LVLMs) have demonstrated their incredible capability in image understanding and response generation. However, this rich visual interaction also makes LVLMs vulnerable to adversarial examples. In this paper, we formulate a novel and practical targeted attack scenario that the adversary can only know the vision encoder of the victim LVLM, without the knowledge of its prompts (which are often proprietary for service providers and not publicly available) and its underlying large language model (LLM). This practical setting poses challenges to the cross-prompt and cross-model transferability of targeted adversarial attack, which aims to confuse the LVLM to output a response that is semantically similar to the attacker's chosen target text. To this end, we propose an instruction-tuned targeted attack (dubbed \textsc{InstructTA}) to deliver the targeted adversarial attack on LVLMs with high transferability. Initially, we utilize a public text-to-image generative model to "reverse" the target response into a target image, and employ GPT-4 to infer a reasonable instruction $\boldsymbol{p}^\prime$ from the target response. We then form a local surrogate model (sharing the same vision encoder with the victim LVLM) to extract instruction-aware features of an adversarial image example and the target image, and minimize the distance between these two features to optimize the adversarial example. To further improve the transferability with instruction tuning, we augment the instruction $\boldsymbol{p}^\prime$ with instructions paraphrased from GPT-4. Extensive experiments demonstrate the superiority of our proposed method in targeted attack performance and transferability. The code is available at https://github.com/xunguangwang/InstructTA.

arXiv:2303.17651

Self-Refine: Iterative Refinement with Self-Feedback

Like humans, large language models (LLMs) do not always generate the best output on their first try. Motivated by how humans refine their written text, we introduce Self-Refine, an approach for improving initial outputs from LLMs through iterative feedback and refinement. The main idea is to generate an initial output using an LLMs; then, the same LLMs provides feedback for its output and uses it to refine itself, iteratively. Self-Refine does not require any supervised training data, additional training, or reinforcement learning, and instead uses a single LLM as the generator, refiner, and feedback provider. We evaluate Self-Refine across 7 diverse tasks, ranging from dialog response generation to mathematical reasoning, using state-of-the-art (GPT-3.5, ChatGPT, and GPT-4) LLMs. Across all evaluated tasks, outputs generated with Self-Refine are preferred by humans and automatic metrics over those generated with the same LLM using conventional one-step generation, improving by ~20% absolute on average in task performance. Our work demonstrates that even state-of-the-art LLMs like GPT-4 can be further improved at test time using our simple, standalone approach.

arXiv:2305.16934공식 코드 · Papers with Code

On Evaluating Adversarial Robustness of Large Vision-Language Models

Large vision-language models (VLMs) such as GPT-4 have achieved unprecedented performance in response generation, especially with visual inputs, enabling more creative and adaptable interaction than large language models such as ChatGPT. Nonetheless, multimodal generation exacerbates safety concerns, since adversaries may successfully evade the entire system by subtly manipulating the most vulnerable modality (e.g., vision). To this end, we propose evaluating the robustness of open-source large VLMs in the most realistic and high-risk setting, where adversaries have only black-box system access and seek to deceive the model into returning the targeted responses. In particular, we first craft targeted adversarial examples against pretrained models such as CLIP and BLIP, and then transfer these adversarial examples to other VLMs such as MiniGPT-4, LLaVA, UniDiffuser, BLIP-2, and Img2Prompt. In addition, we observe that black-box queries on these VLMs can further improve the effectiveness of targeted evasion, resulting in a surprisingly high success rate for generating targeted responses. Our findings provide a quantitative understanding regarding the adversarial vulnerability of large VLMs and call for a more thorough examination of their potential security flaws before deployment in practice. Code is at https://github.com/yunqing-me/AttackVLM.

arXiv:2311.16137

A Graph-to-Text Approach to Knowledge-Grounded Response Generation in Human-Robot Interaction

Knowledge graphs are often used to represent structured information in a flexible and efficient manner, but their use in situated dialogue remains under-explored. This paper presents a novel conversational model for human--robot interaction that rests upon a graph-based representation of the dialogue state. The knowledge graph representing the dialogue state is continuously updated with new observations from the robot sensors, including linguistic, situated and multimodal inputs, and is further enriched by other modules, in particular for spatial understanding. The neural conversational model employed to respond to user utterances relies on a simple but effective graph-to-text mechanism that traverses the dialogue state graph and converts the traversals into a natural language form. This conversion of the state graph into text is performed using a set of parameterized functions, and the values for those parameters are optimized based on a small set of Wizard-of-Oz interactions. After this conversion, the text representation of the dialogue state graph is included as part of the prompt of a large language model used to decode the agent response. The proposed approach is empirically evaluated through a user study with a humanoid robot that acts as conversation partner to evaluate the impact of the graph-to-text mechanism on the response generation. After moving a robot along a tour of an indoor environment, participants interacted with the robot using spoken dialogue and evaluated how well the robot was able to answer questions about what the robot observed during the tour. User scores show a statistically significant improvement in the perceived factuality of the robot responses when the graph-to-text approach is employed, compared to a baseline using inputs structured as semantic triples.

arXiv:2311.15016

E-CORE: Emotion Correlation Enhanced Empathetic Dialogue Generation

Achieving empathy is a crucial step toward humanized dialogue systems. Current approaches for empathetic dialogue generation mainly perceive an emotional label to generate an empathetic response conditioned on it, which simply treat emotions independently, but ignore the intrinsic emotion correlation in dialogues, resulting in inaccurate emotion perception and unsuitable response generation. In this paper, we propose a novel emotion correlation enhanced empathetic dialogue generation framework, which comprehensively realizes emotion correlation learning, utilization, and supervising. Specifically, a multi-resolution emotion graph is devised to capture context-based emotion interactions from different resolutions, further modeling emotion correlation. Then we propose an emotion correlation enhanced decoder, with a novel correlation-aware aggregation and soft/hard strategy, respectively improving the emotion perception and response generation. Experimental results on the benchmark dataset demonstrate the superiority of our model in both empathetic perception and expression.

arXiv:2311.10587공식 코드 · Papers with Code

Countering Misinformation via Emotional Response Generation

The proliferation of misinformation on social media platforms (SMPs) poses a significant danger to public health, social cohesion and ultimately democracy. Previous research has shown how social correction can be an effective way to curb misinformation, by engaging directly in a constructive dialogue with users who spread -- often in good faith -- misleading messages. Although professional fact-checkers are crucial to debunking viral claims, they usually do not engage in conversations on social media. Thereby, significant effort has been made to automate the use of fact-checker material in social correction; however, no previous work has tried to integrate it with the style and pragmatics that are commonly employed in social media communication. To fill this gap, we present VerMouth, the first large-scale dataset comprising roughly 12 thousand claim-response pairs (linked to debunking articles), accounting for both SMP-style and basic emotions, two factors which have a significant role in misinformation credibility and spreading. To collect this dataset we used a technique based on an author-reviewer pipeline, which efficiently combines LLMs and human annotators to obtain high-quality data. We also provide comprehensive experiments showing how models trained on our proposed dataset have significant improvements in terms of output quality and generalization capabilities.

arXiv:2311.09832

WatME: Towards Lossless Watermarking Through Lexical Redundancy

Text watermarking has emerged as a pivotal technique for identifying machine-generated text. However, existing methods often rely on arbitrary vocabulary partitioning during decoding to embed watermarks, which compromises the availability of suitable tokens and significantly degrades the quality of responses. This study assesses the impact of watermarking on different capabilities of large language models (LLMs) from a cognitive science lens. Our finding highlights a significant disparity; knowledge recall and logical reasoning are more adversely affected than language generation. These results suggest a more profound effect of watermarking on LLMs than previously understood. To address these challenges, we introduce Watermarking with Mutual Exclusion (WatME), a novel approach leveraging linguistic prior knowledge of inherent lexical redundancy in LLM vocabularies to seamlessly integrate watermarks. Specifically, WatME dynamically optimizes token usage during the decoding process by applying a mutually exclusive rule to the identified lexical redundancies. This strategy effectively prevents the unavailability of appropriate tokens and preserves the expressive power of LLMs. We provide both theoretical analysis and empirical evidence showing that WatME effectively preserves the diverse capabilities of LLMs while ensuring watermark detectability.

arXiv:2311.09136공식 코드 · Papers with Code

Rescue: Ranking LLM Responses with Partial Ordering to Improve Response Generation

Customizing LLMs for a specific task involves separating high-quality responses from lower-quality ones. This skill can be developed using supervised fine-tuning with extensive human preference data. However, obtaining a large volume of expert-annotated data is costly for most tasks. In this paper, we explore a novel method to optimize LLMs using ranking metrics. This method trains the model to prioritize the best responses from a pool of candidates created for a particular task. Rather than a traditional full ordering, we advocate for a partial ordering, as achieving consensus on the perfect order of candidate responses can be challenging. Our partial ordering is more robust, less sensitive to noise, and can be achieved with limited human annotations or through heuristic methods. We test our system's improved response generation ability using benchmark datasets, including textual entailment and multi-document question answering. We conduct ablation studies to understand crucial factors, such as how to gather candidate responses for a specific task, determine their most suitable order, and balance supervised fine-tuning with ranking metrics. Our approach, named Rescue, offers a promising avenue for enhancing the response generation and task accuracy of LLMs.