Search

The emergence of text-to-image synthesis (TIS) models has significantly influenced digital image creation by producing high-quality visuals from written descriptions. Yet these models heavily rely on the quality and specificity of textual prompts, posing a challenge for novice users who may not be familiar with TIS-model-preferred prompt writing. Existing solutions relieve this via automatic model-preferred prompt generation from user queries. However, this single-turn manner suffers from limited user-centricity in terms of result interpretability and user interactivity. To address these issues, we propose DialPrompt, a multi-turn dialogue-based TIS prompt generation model that emphasises user-centricity. DialPrompt is designed to follow a multi-turn guidance workflow, where in each round of dialogue the model queries user with their preferences on possible optimization dimensions before generating the final TIS prompt. To achieve this, we mined 15 essential dimensions for high-quality prompts from advanced users and curated a multi-turn dataset. Through training on this dataset, DialPrompt can improve interpretability by allowing users to understand the correlation between specific phrases and image attributes. Additionally, it enables greater user control and engagement in the prompt generation process, leading to more personalized and visually satisfying outputs. Experiments indicate that DialPrompt achieves a competitive result in the quality of synthesized images, outperforming existing prompt engineering approaches by 5.7%. Furthermore, in our user evaluation, DialPrompt outperforms existing approaches by 46.5% in user-centricity score and is rated 7.9/10 by 19 human reviewers.

In modern online platforms, incentives are essential factors that enhance user engagement and increase platform revenue. Over recent years, uplift modeling has been introduced as a strategic approach to assign incentives to individual customers. Especially in many real-world applications, online platforms can only incentivize customers with specific budget constraints. This problem can be reformulated as the multi-choice knapsack problem. This optimization aims to select the optimal incentive for each customer to maximize the return on investment. Recent works in this field frequently tackle the budget allocation problem using a two-stage approach. However, this solution is confronted with the following challenges: (1) The causal inference methods often ignore the domain knowledge in online marketing, where the expected response curve of a customer should be monotonic and smooth as the incentive increases. (2) An optimality gap between the two stages results in inferior sub-optimal allocation performance due to the loss of the incentive recommendation information for the uplift prediction under the limited budget constraint. To address these challenges, we propose a novel End-to-End Cost-Effective Incentive Recommendation (E3IR) model under budget constraints. Specifically, our methods consist of two modules, i.e., the uplift prediction module and the differentiable allocation module. In the uplift prediction module, we construct prediction heads to capture the incremental improvement between adjacent treatments with the marketing domain constraints (i.e., monotonic and smooth). We incorporate integer linear programming (ILP) as a differentiable layer input in the allocation module. Furthermore, we conduct extensive experiments on public and real product datasets, demonstrating that our E3IR improves allocation performance compared to existing two-stage approaches., Comment: Accepted by RecSys 2024

We study the capability of Video-Language (VidL) models in understanding compositions between objects, attributes, actions and their relations. Composition understanding becomes particularly challenging for video data since the compositional relations rapidly change over time in videos. We first build a benchmark named AARO to evaluate composition understanding related to actions on top of spatial concepts. The benchmark is constructed by generating negative texts with incorrect action descriptions for a given video and the model is expected to pair a positive text with its corresponding video. Furthermore, we propose a training method called NAVERO which utilizes video-text data augmented with negative texts to enhance composition understanding. We also develop a negative-augmented visual-language matching loss which is used explicitly to benefit from the generated negative text. We compare NAVERO with other state-of-the-art methods in terms of compositional understanding as well as video-text retrieval performance. NAVERO achieves significant improvement over other methods for both video-language and image-language composition understanding, while maintaining strong performance on traditional text-video retrieval tasks.

Aiming to generalize the label knowledge from a source domain with continuous outputs to an unlabeled target domain, Domain Adaptation Regression (DAR) is developed for complex practical learning problems. However, due to the continuity problem in regression, existing conditional distribution alignment theory and methods with discrete prior, which are proven to be effective in classification settings, are no longer applicable. In this work, focusing on the feasibility problems in DAR, we establish the sufficiency theory for the regression model, which shows the generalization error can be sufficiently dominated by the cross-domain conditional discrepancy. Further, to characterize conditional discrepancy with continuous conditioning variable, a novel Conditional Operator Discrepancy (COD) is proposed, which admits the metric property on conditional distributions via the kernel embedding theory. Finally, to minimize the discrepancy, a COD-based conditional invariant representation learning model is proposed, and the reformulation is derived to show that reasonable modifications on moment statistics can further improve the discriminability of the adaptation model. Extensive experiments on standard DAR datasets verify the validity of theoretical results and the superiority over SOTA DAR methods., Comment: Accepted to ECCV 2024 (oral)

In sectors such as finance and healthcare, where data governance is subject to rigorous regulatory requirements, the exchange and utilization of data are particularly challenging. Federated Learning (FL) has risen as a pioneering distributed machine learning paradigm that enables collaborative model training across multiple institutions while maintaining data decentralization. Despite its advantages, FL is vulnerable to adversarial threats, particularly poisoning attacks during model aggregation, a process typically managed by a central server. However, in these systems, neural network models still possess the capacity to inadvertently memorize and potentially expose individual training instances. This presents a significant privacy risk, as attackers could reconstruct private data by leveraging the information contained in the model itself. Existing solutions fall short of providing a viable, privacy-preserving BRFL system that is both completely secure against information leakage and computationally efficient. To address these concerns, we propose Lancelot, an innovative and computationally efficient BRFL framework that employs fully homomorphic encryption (FHE) to safeguard against malicious client activities while preserving data privacy. Our extensive testing, which includes medical imaging diagnostics and widely-used public image datasets, demonstrates that Lancelot significantly outperforms existing methods, offering more than a twenty-fold increase in processing speed, all while maintaining data privacy., Comment: 26 pages

Seismic tomography is a crucial technique used to image subsurface structures at various scales, accomplished by solving a nonlinear and nonunique inverse problem. It is therefore important to quantify velocity model uncertainties for accurate earthquake locations and geological interpretations. Monte Carlo sampling techniques are usually used for this purpose, but those methods are computationally intensive, especially for large datasets or high-dimensional parameter spaces. In comparison, Bayesian variational inference provides a more efficient alternative by delivering probabilistic solutions through optimization. The method has been proven to be efficient in 2D tomographic problems. In this study, we apply variational inference to solve 3D double-difference (DD) seismic tomographic system using both absolute and differential travel time data. Synthetic tests demonstrate that the new method can produce more accurate velocity models than the original DD tomography method by avoiding regularization constraints, and at the same time provides more reliable uncertainty estimates. Compared to traditional checkerboard resolution tests, the resulting uncertainty estimates measure more accurately the reliability of the solution. We further apply the new method to data recorded by a local dense seismic array around the San Andreas Fault Observatory at Depth (SAFOD) site along the San Andreas Fault (SAF) at Parkfield. Similar to other researches, the obtained velocity models show significant velocity contrasts across the fault. More importantly, the new method produces velocity uncertainties of less than 0.34 km/s for Vp and 0.23 km/s for Vs. We therefore conclude that variational inference provides a powerful and efficient tool for solving 3D seismic tomographic problems and quantifying model uncertainties.

Large language models (LLMs) as autonomous agents offer a novel avenue for tackling real-world challenges through a knowledge-driven manner. These LLM-enhanced methodologies excel in generalization and interpretability. However, the complexity of driving tasks often necessitates the collaboration of multiple, heterogeneous agents, underscoring the need for such LLM-driven agents to engage in cooperative knowledge sharing and cognitive synergy. Despite the promise of LLMs, current applications predominantly center around single agent scenarios. To broaden the horizons of knowledge-driven strategies and bolster the generalization capabilities of autonomous agents, we propose the KoMA framework consisting of multi-agent interaction, multi-step planning, shared-memory, and ranking-based reflection modules to enhance multi-agents' decision-making in complex driving scenarios. Based on the framework's generated text descriptions of driving scenarios, the multi-agent interaction module enables LLM agents to analyze and infer the intentions of surrounding vehicles, akin to human cognition. The multi-step planning module enables LLM agents to analyze and obtain final action decisions layer by layer to ensure consistent goals for short-term action decisions. The shared memory module can accumulate collective experience to make superior decisions, and the ranking-based reflection module can evaluate and improve agent behavior with the aim of enhancing driving safety and efficiency. The KoMA framework not only enhances the robustness and adaptability of autonomous driving agents but also significantly elevates their generalization capabilities across diverse scenarios. Empirical results demonstrate the superiority of our approach over traditional methods, particularly in its ability to handle complex, unpredictable driving environments without extensive retraining., Comment: 13 pages, 18 figures

The emotion recognition has attracted more attention in recent decades. Although significant progress has been made in the recognition technology of the seven basic emotions, existing methods are still hard to tackle compound emotion recognition that occurred commonly in practical application. This article introduces our achievements in the 7th Field Emotion Behavior Analysis (ABAW) competition. In the competition, we selected pre trained ResNet18 and Transformer, which have been widely validated, as the basic network framework. Considering the continuity of emotions over time, we propose a time pyramid structure network for frame level emotion prediction. Furthermore. At the same time, in order to address the lack of data in composite emotion recognition, we utilized fine-grained labels from the DFEW database to construct training data for emotion categories in competitions. Taking into account the characteristics of valence arousal of various complex emotions, we constructed a classification framework from coarse to fine in the label space., Comment: draft for abaw7

Recently, video diffusion models (VDMs) have garnered significant attention due to their notable advancements in generating coherent and realistic video content. However, processing multiple frame features concurrently, coupled with the considerable model size, results in high latency and extensive memory consumption, hindering their broader application. Post-training quantization (PTQ) is an effective technique to reduce memory footprint and improve computational efficiency. Unlike image diffusion, we observe that the temporal features, which are integrated into all frame features, exhibit pronounced skewness. Furthermore, we investigate significant inter-channel disparities and asymmetries in the activation of video diffusion models, resulting in low coverage of quantization levels by individual channels and increasing the challenge of quantization. To address these issues, we introduce the first PTQ strategy tailored for video diffusion models, dubbed QVD. Specifically, we propose the High Temporal Discriminability Quantization (HTDQ) method, designed for temporal features, which retains the high discriminability of quantized features, providing precise temporal guidance for all video frames. In addition, we present the Scattered Channel Range Integration (SCRI) method which aims to improve the coverage of quantization levels across individual channels. Experimental validations across various models, datasets, and bit-width settings demonstrate the effectiveness of our QVD in terms of diverse metrics. In particular, we achieve near-lossless performance degradation on W8A8, outperforming the current methods by 205.12 in FVD., Comment: accepted by ACMMM2024

In the rapidly evolving field of natural language processing, dialogue systems primarily employ a single-step dialogue paradigm. Although this paradigm is efficient, it lacks the depth and fluidity of human interactions and does not appear natural. We introduce a novel \textbf{Step}-by-Step Dialogue Paradigm (Stephanie), designed to mimic the ongoing dynamic nature of human conversations. By employing a dual learning strategy and a further-split post-editing method, we generated and utilized a high-quality step-by-step dialogue dataset to fine-tune existing large language models, enabling them to perform step-by-step dialogues. We thoroughly present Stephanie. Tailored automatic and human evaluations are conducted to assess its effectiveness compared to the traditional single-step dialogue paradigm. We will release code, Stephanie datasets, and Stephanie LLMs to facilitate the future of chatbot eras.

Abstractive Speech Summarization (SSum) aims to generate human-like text summaries from spoken content. It encounters difficulties in handling long speech input and capturing the intricate cross-modal mapping between long speech inputs and short text summaries. Research on large language models (LLMs) and multimodal information fusion has provided new insights for addressing these challenges. In this paper, we propose an end-to-end SSum model that utilizes Q-Former as a connector for the audio-text modality and employs LLMs to generate text summaries directly from speech features. We adopt a multi-stage training approach that includes LLM based ASR and Text Summarization (TSum) tasks as auxiliary tasks. ASR tasks are used to align feature spaces and enhance the LLM's ability to handle longer speech. Then, we utilize a curriculum learning strategy to facilitate the model's transition from TSum to SSum. Finally, our model achieves competitive performance on the How-2 dataset., Comment: InterSpeech 2024

The inclusive production of pseudoscalar heavy quarkoniua ($\eta_c,~\eta_b$ and $B_c$) via photon-photon fusion in heavy ion ultraperipheral collision (UPC) are calculated to QCD next-to-leading order in the framework of non-relativistic QCD (NRQCD). The total cross section of $\eta_c$ produced in Pb-Pb UPC is 194 $\mathrm{nb}$ and 1275 $\mathrm{nb}$ at nucleon-nucleon c.m. energies $\sqrt{S_{\mathrm{NN}}}=$ 5.52 TeV and 39.4 TeV, respectively. The cross sections for $\eta_b$ and $B_c$ mesons are more than two to three orders of magnitude smaller. We make a detailed phenomenological analysis on the $\eta_c$ production; the uncertainties caused by the renormalization scale and the charm quark mass, the cross sections in other ultraperipheral nucleon-nucleon colliding systems, and the transverse momentum distribution are discussed. At the coming HL-LHC and future FCC, the heavy ion UPC opens another door of the study on the production of heavy quarkonium., Comment: 17 pages, 1 figures, 4 tables

Large Multimodal Models (LMMs) have achieved great success recently, demonstrating a strong capability to understand multimodal information and to interact with human users. Despite the progress made, the challenge of detecting high-risk interactions in multimodal settings, and in particular in speech modality, remains largely unexplored. Conventional research on risk for speech modality primarily emphasises the content (e.g., what is captured as transcription). However, in speech-based interactions, paralinguistic cues in audio can significantly alter the intended meaning behind utterances. In this work, we propose a speech-specific risk taxonomy, covering 8 risk categories under hostility (malicious sarcasm and threats), malicious imitation (age, gender, ethnicity), and stereotypical biases (age, gender, ethnicity). Based on the taxonomy, we create a small-scale dataset for evaluating current LMMs capability in detecting these categories of risk. We observe even the latest models remain ineffective to detect various paralinguistic-specific risks in speech (e.g., Gemini 1.5 Pro is performing only slightly above random baseline). Warning: this paper contains biased and offensive examples.

Vision-language models (VLMs) have been applied to robot task planning problems, where the robot receives a task in natural language and generates plans based on visual inputs. While current VLMs have demonstrated strong vision-language understanding capabilities, their performance is still far from being satisfactory in planning tasks. At the same time, although classical task planners, such as PDDL-based, are strong in planning for long-horizon tasks, they do not work well in open worlds where unforeseen situations are common. In this paper, we propose a novel task planning and execution framework, called DKPROMPT, which automates VLM prompting using domain knowledge in PDDL for classical planning in open worlds. Results from quantitative experiments show that DKPROMPT outperforms classical planning, pure VLM-based and a few other competitive baselines in task completion rate.

The increasing rate of road accidents worldwide results not only in significant loss of life but also imposes billions financial burdens on societies. Current research in traffic crash frequency modeling and analysis has predominantly approached the problem as classification tasks, focusing mainly on learning-based classification or ensemble learning methods. These approaches often overlook the intricate relationships among the complex infrastructure, environmental, human and contextual factors related to traffic crashes and risky situations. In contrast, we initially propose a large-scale traffic crash language dataset, named CrashEvent, summarizing 19,340 real-world crash reports and incorporating infrastructure data, environmental and traffic textual and visual information in Washington State. Leveraging this rich dataset, we further formulate the crash event feature learning as a novel text reasoning problem and further fine-tune various large language models (LLMs) to predict detailed accident outcomes, such as crash types, severity and number of injuries, based on contextual and environmental factors. The proposed model, CrashLLM, distinguishes itself from existing solutions by leveraging the inherent text reasoning capabilities of LLMs to parse and learn from complex, unstructured data, thereby enabling a more nuanced analysis of contributing factors. Our experiments results shows that our LLM-based approach not only predicts the severity of accidents but also classifies different types of accidents and predicts injury outcomes, all with averaged F1 score boosted from 34.9% to 53.8%. Furthermore, CrashLLM can provide valuable insights for numerous open-world what-if situational-awareness traffic safety analyses with learned reasoning features, which existing models cannot offer. We make our benchmark, datasets, and model public available for further exploration.

The fallback disk model is widely accepted to explain long-period neutron stars (NSs) which can't be simulated by magnetic dipole radiation. However, no confirmed detection of disk was found from the newly discovered long period pulsars GLEAM-X 162759.5-523504.3, GPM J1839-10 and the known slowest isolated NSs 1E 161348-5055. This might be that the disks have either been in noninteracting/inactive state where its emission is too weak to be detected or have been disrupted. In this work, we conduct simulations to examine the lifetime of supernova fallback disks around isolated neutron stars. We assume that the disk's mass varies in a self-similar way and its interaction with the NS occurs only in interacting/active state. Our results reveal that nearly all the interacting lifetimes for the disk are shorter than 0.1 Myr while the existence lifetimes are considerably longer., Comment: Accepted in ApJ, comments are welcome

Compared to traditional sentiment analysis, which only considers text, multimodal sentiment analysis needs to consider emotional signals from multimodal sources simultaneously and is therefore more consistent with the way how humans process sentiment in real-world scenarios. It involves processing emotional information from various sources such as natural language, images, videos, audio, physiological signals, etc. However, although other modalities also contain diverse emotional cues, natural language usually contains richer contextual information and therefore always occupies a crucial position in multimodal sentiment analysis. The emergence of ChatGPT has opened up immense potential for applying large language models (LLMs) to text-centric multimodal tasks. However, it is still unclear how existing LLMs can adapt better to text-centric multimodal sentiment analysis tasks. This survey aims to (1) present a comprehensive review of recent research in text-centric multimodal sentiment analysis tasks, (2) examine the potential of LLMs for text-centric multimodal sentiment analysis, outlining their approaches, advantages, and limitations, (3) summarize the application scenarios of LLM-based multimodal sentiment analysis technology, and (4) explore the challenges and potential research directions for multimodal sentiment analysis in the future., Comment: arXiv admin note: text overlap with arXiv:2210.14556 by other authors

Despite recent improvements in End-to-End Automatic Speech Recognition (E2E ASR) systems, the performance can degrade due to vocal characteristic mismatches between training and testing data, particularly with limited target speaker adaptation data. We propose a novel speaker adaptation approach Speaker-Smoothed kNN that leverages k-Nearest Neighbors (kNN) retrieval techniques to improve model output by finding correctly pronounced tokens from its pre-built datastore during the decoding phase. Moreover, we utilize x-vector to dynamically adjust kNN interpolation parameters for data sparsity issue. This approach was validated using KeSpeech and MagicData corpora under in-domain and all-domain settings. Our method consistently performs comparably to fine-tuning without the associated performance degradation during speaker changes. Furthermore, in the all-domain setting, our method achieves state-of-the-art results, reducing the CER in both single speaker and multi-speaker test scenarios., Comment: Accepted to Interspeech 2024

We investigate the waveforms and quasi-normal modes around Schwarzschild traversable wormholes under different field perturbations, including the scalar field perturbation, the electromagnetic (vector) field perturbation and the gravitational (tensor) field perturbation. By comparing Schwarzschild traversable wormholes with Schwarzschild black holes, we find some unique properties for the former. At first, the perturbation waveform includes echoes and damping oscillations around Schwarzschild traversable wormholes, while it has only the damping waveform around Schwarzschild black holes. Secondly, the difference between adjacent peaks of echoes varies with the mass parameter and the defect parameter in the waveform around Schwarzschild traversable wormholes, while it always keeps constant around Schwarzschild black holes. Thirdly, the ordinary isospectrality between the odd and even parities no longer exists in the quasi-normal modes of gravitational perturbations around Schwarzschild traversable wormholes, but an alternative isospectrality appears. According to these properties, we summarize a scenario for estimating the mass parameter and the defect parameter of Schwarzschild traversable wormholes through the waveforms and quasi-normal modes. Our analyses provide a more profound comprehension of the inherent characteristics of Schwarzschild traversable wormholes., Comment: v1: 28 pages, 12 figures, 9 tables

In transportation networks, intersections pose significant risks of collisions due to conflicting movements of vehicles approaching from different directions. To address this issue, various tools can exert influence on traffic safety both directly and indirectly. This study focuses on investigating the impact of adaptive signal control and connected and automated vehicles (CAVs) on intersection safety using a deep reinforcement learning approach. The objective is to assess the individual and combined effects of CAVs and adaptive traffic signal control on traffic safety, considering rear-end and crossing conflicts. The study employs a Deep Q Network (DQN) to regulate traffic signals and driving behaviors of both CAVs and Human Drive Vehicles (HDVs), and uses Time To Collision (TTC) metric to evaluate safety. The findings demonstrate a significant reduction in rear-end and crossing conflicts through the combined implementation of CAVs and DQNs-based traffic signal control. Additionally, the long-term positive effects of CAVs on safety are similar to the short-term effects of combined CAVs and DQNs-based traffic signal control. Overall, the study emphasizes the potential benefits of integrating CAVs and adaptive traffic signal control approaches in order to enhance traffic safety. The findings of this study could provide valuable insights for city officials and transportation authorities in developing effective strategies to improve safety at signalized intersections., Comment: TRB 103nd Annual Meeting

This paper studies observability for non-uniform hypergraphs with inputs and outputs. To capture higher-order interactions, we define a canonical non-homogeneous dynamical system with nonlinear outputs on hypergraphs. We then construct algebraic necessary and sufficient conditions based on polynomial ideals and varieties for global observability at an initial state of hypergraphs. An example is given to illustrate the proposed criteria for observability. Further, necessary and sufficient conditions for local observability are derived based on rank conditions of observability matrices, which provide a framework to study local observability for non-uniform hypergraphs. Finally, the similarity of observability for hypergraphs is proposed using similarity of tensors, which reveals the relation of observability between two hypergraphs and helps to check the observability intuitively.

As a promising individualized treatment effect (ITE) estimation method, counterfactual regression (CFR) maps individuals' covariates to a latent space and predicts their counterfactual outcomes. However, the selection bias between control and treatment groups often imbalances the two groups' latent distributions and negatively impacts this method's performance. In this study, we revisit counterfactual regression through the lens of information bottleneck and propose a novel learning paradigm called Gromov-Wasserstein information bottleneck (GWIB). In this paradigm, we learn CFR by maximizing the mutual information between covariates' latent representations and outcomes while penalizing the kernelized mutual information between the latent representations and the covariates. We demonstrate that the upper bound of the penalty term can be implemented as a new regularizer consisting of $i)$ the fused Gromov-Wasserstein distance between the latent representations of different groups and $ii)$ the gap between the transport cost generated by the model and the cross-group Gromov-Wasserstein distance between the latent representations and the covariates. GWIB effectively learns the CFR model through alternating optimization, suppressing selection bias while avoiding trivial latent distributions. Experiments on ITE estimation tasks show that GWIB consistently outperforms state-of-the-art CFR methods. To promote the research community, we release our project at https://github.com/peteryang1031/Causal-GWIB., Comment: 19 pages

The scarcity of non-English data limits the development of non-English large language models (LLMs). Transforming English-centric LLMs to non-English has been identified as an effective and resource-efficient method. Previous works start from base LLMs and perform knowledge distillation (KD) with data generated by stronger LLMs, e.g. GPT-4. Compared to base LLMs, chat LLMs are further optimized for advanced abilities, e.g. multi-turn conversation and human preference alignment, and thus more powerful in both helpfulness and safety. However, transforming a chat LLM involves two critical issues: (1) How can we effectively transfer advanced abilities without their supervised data? (2) How can we prevent the original knowledge from catastrophic forgetting during transformation? We target these issues by introducing a simple framework called TransLLM. For the first issue, TransLLM divides the transfer problem into some common sub-tasks with the translation chain-of-thought, which uses the translation as the bridge between English and non-English step-by-step. We further enhance the performance of sub-tasks with publicly available data. For the second issue, we propose a method comprising two synergistic components: low-rank adaptation for training to maintain the original LLM parameters, and recovery KD, which utilizes data generated by the chat LLM itself to recover the original knowledge from the frozen parameters. In the experiments, we transform the LLaMA-2-chat-7B to the Thai language. Our method, using only single-turn data, outperforms strong baselines and ChatGPT on multi-turn benchmark MT-bench. Furthermore, our method, without safety data, rejects more harmful queries of safety benchmark AdvBench than both ChatGPT and GPT-4.

One critical prerequisite for faithful text-to-image generation is the accurate understanding of text inputs. Existing methods leverage the text encoder of the CLIP model to represent input prompts. However, the pre-trained CLIP model can merely encode English with a maximum token length of 77. Moreover, the model capacity of the text encoder from CLIP is relatively limited compared to Large Language Models (LLMs), which offer multilingual input, accommodate longer context, and achieve superior text representation. In this paper, we investigate LLMs as the text encoder to improve the language understanding in text-to-image generation. Unfortunately, training text-to-image generative model with LLMs from scratch demands significant computational resources and data. To this end, we introduce a three-stage training pipeline that effectively and efficiently integrates the existing text-to-image model with LLMs. Specifically, we propose a lightweight adapter that enables fast training of the text-to-image model using the textual representations from LLMs. Extensive experiments demonstrate that our model supports not only multilingual but also longer input context with superior image generation quality., Comment: To appear in ECCV-2024, Project page: https://llm-conditioned-diffusion.github.io/