The Telephone Game: Evaluating Semantic Drift in Unified Models

• URL: http://arxiv.org/abs/2509.04438v2
• Date: Mon, 06 Oct 2025 17:49:39 GMT
• Title: The Telephone Game: Evaluating Semantic Drift in Unified Models
• Authors: Sabbir Mollah, Rohit Gupta, Sirnam Swetha, Qingyang Liu, Ahnaf Munir, Mubarak Shah,
• Abstract summary: Employing a single, unified model (UM) for both visual understanding (image-to-text: I2T) and visual generation (text-to-image: T2I) has opened a new direction in Visual Language Model (VLM) research.<n>Existing evaluation benchmarks consider these capabilities in isolation: FID and GenEval for T2I, and benchmarks such as MME, MMBench for I2T.<n>These isolated single-pass metrics do not reveal cross-consistency: whether a model that "understands" a concept can also "render" it, nor whether semantic meaning
• Score: 41.650904633974584
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Employing a single, unified model (UM) for both visual understanding (image-to-text: I2T) and visual generation (text-to-image: T2I) has opened a new direction in Visual Language Model (VLM) research. While UMs can also support broader unimodal tasks (e.g., text-to-text, image-to-image), we focus on the core cross-modal pair T2I and I2T. Existing evaluation benchmarks consider these capabilities in isolation: FID and GenEval for T2I, and benchmarks such as MME, MMBench for I2T. These isolated single-pass metrics do not reveal cross-consistency: whether a model that "understands" a concept can also "render" it, nor whether semantic meaning is preserved when cycling between image and text modalities. To address this, we introduce the Semantic Drift Protocol (SDP) for Unified Models, a cyclic evaluation protocol that alternates I2T and T2I over multiple generations to quantify semantic drift. We propose two metrics: (i) Mean Cumulative Drift (MCD), an embedding-based measure of overall semantic drift; and (ii) Multi-Generation GenEval (MGG), an object-level compliance score extending GenEval. To assess generalization beyond COCO dataset, which is widely used in training; we create a new benchmark Nocaps+Docci400, sampled from NoCaps and DOCCI and evaluated on seven recent models. SDP reveals substantial variation in cross-modal stability: some models like BAGEL maintain semantic meaning over many alternations, whereas others like VILA-U drift quickly despite strong single-pass scores. Our results highlight SDP as a necessary complement to standard I2T and T2I evaluations. Code is available at https://github.com/mollahsabbir/Semantic-Drift-in-Unified-Models

Related papers

• Exploring a Unified Vision-Centric Contrastive Alternatives on Multi-Modal Web Documents [99.62178668680578]
  We propose Vision-Centric Contrastive Learning (VC2L), a unified framework that models text, images, and their combinations using a single vision transformer.<n> VC2L operates entirely in pixel space by rendering all inputs, whether textual, visual, or combined, as images.<n>To capture complex cross-modal relationships in web documents, VC2L employs a snippet-level contrastive learning objective that aligns consecutive multimodal segments.
  arXiv  Detail & Related papers  (2025-10-21T14:59:29Z)
• UniGenBench++: A Unified Semantic Evaluation Benchmark for Text-to-Image Generation [40.644151228285246]
  We introduce UniGenBench++, a unified semantic assessment benchmark for text-to-image generation.<n>It comprises 600 prompts organized hierarchically to ensure both coverage and efficiency.<n>It provides both English and Chinese versions of each prompt in short and long forms.
  arXiv  Detail & Related papers  (2025-10-21T14:56:46Z)
• Test-Time Consistency in Vision Language Models [26.475993408532304]
  Vision-Language Models (VLMs) have achieved impressive performance across a wide range of multimodal tasks.<n>Recent benchmarks, such as MM-R3, highlight that even state-of-the-art VLMs can produce divergent predictions across semantically equivalent inputs.<n>We propose a simple and effective test-time consistency framework that enhances semantic consistency without supervised re-training.
  arXiv  Detail & Related papers  (2025-06-27T17:09:44Z)
• TIIF-Bench: How Does Your T2I Model Follow Your Instructions? [7.13169573900556]
  We present TIIF-Bench (Text-to-Image Instruction Following Benchmark), aiming to systematically assess T2I models' ability in interpreting and following intricate textual instructions.<n> TIIF-Bench comprises a set of 5000 prompts organized along multiple dimensions, which are categorized into three levels of difficulties and complexities.<n>Two critical attributes, i.e. text rendering and style control, are introduced to evaluate the precision of text synthesis and the aesthetic coherence of T2I models.
  arXiv  Detail & Related papers  (2025-06-02T18:44:07Z)
• Multi-Modal Language Models as Text-to-Image Model Evaluators [16.675735328424786]
  Multimodal Text-to-Image Eval (MT2IE) is an evaluation framework that iteratively generates prompts for evaluation.<n>We show that MT2IE's prompt-generation consistency scores have higher correlation with human judgment than scores previously introduced in the literature.
  arXiv  Detail & Related papers  (2025-05-01T17:47:55Z)
• UniCMs: A Unified Consistency Model For Efficient Multimodal Generation and Understanding [12.34529497235534]
  consistency models (CMs) have shown promise in efficient generation of both image and text.<n>Key challenge is establishing a unified denoising perspective for both image and text generation.<n>In text-to-image generation, UniCMs outperform SD3 on GenEval, Image Reward, and CLIP Score metrics.<n>In image-to-text generation, UniCMs surpass Show-o on the MMMU benchmark while being $1.5 times$ faster at long-sequence generating speed.
  arXiv  Detail & Related papers  (2025-02-08T02:52:25Z)
• EvalMuse-40K: A Reliable and Fine-Grained Benchmark with Comprehensive Human Annotations for Text-to-Image Generation Model Evaluation [29.176750442205325]
  In this study, we contribute an EvalMuse-40K benchmark, gathering 40K image-text pairs with fine-grained human annotations for image-text alignment-related tasks.<n>We introduce two new methods to evaluate the image-text alignment capabilities of T2I models.
  arXiv  Detail & Related papers  (2024-12-24T04:08:25Z)
• Who Evaluates the Evaluations? Objectively Scoring Text-to-Image Prompt Coherence Metrics with T2IScoreScore (TS2) [62.44395685571094]
  We introduce T2IScoreScore, a curated set of semantic error graphs containing a prompt and a set of increasingly erroneous images. These allow us to rigorously judge whether a given prompt faithfulness metric can correctly order images with respect to their objective error count. We find that the state-of-the-art VLM-based metrics fail to significantly outperform simple (and supposedly worse) feature-based metrics like CLIPScore.
  arXiv  Detail & Related papers  (2024-04-05T17:57:16Z)
• Visual Programming for Text-to-Image Generation and Evaluation [73.12069620086311]
  We propose two novel interpretable/explainable visual programming frameworks for text-to-image (T2I) generation and evaluation. First, we introduce VPGen, an interpretable step-by-step T2I generation framework that decomposes T2I generation into three steps: object/count generation, layout generation, and image generation. Second, we introduce VPEval, an interpretable and explainable evaluation framework for T2I generation based on visual programming.
  arXiv  Detail & Related papers  (2023-05-24T16:42:17Z)
• USER: Unified Semantic Enhancement with Momentum Contrast for Image-Text Retrieval [115.28586222748478]
  Image-Text Retrieval (ITR) aims at searching for the target instances that are semantically relevant to the given query from the other modality. Existing approaches typically suffer from two major limitations.
  arXiv  Detail & Related papers  (2023-01-17T12:42:58Z)
• Joint Spatial-Temporal and Appearance Modeling with Transformer for Multiple Object Tracking [59.79252390626194]
  We propose a novel solution named TransSTAM, which leverages Transformer to model both the appearance features of each object and the spatial-temporal relationships among objects. The proposed method is evaluated on multiple public benchmarks including MOT16, MOT17, and MOT20, and it achieves a clear performance improvement in both IDF1 and HOTA.
  arXiv  Detail & Related papers  (2022-05-31T01:19:18Z)
• Adaptive Context-Aware Multi-Modal Network for Depth Completion [107.15344488719322]
  We propose to adopt the graph propagation to capture the observed spatial contexts. We then apply the attention mechanism on the propagation, which encourages the network to model the contextual information adaptively. Finally, we introduce the symmetric gated fusion strategy to exploit the extracted multi-modal features effectively. Our model, named Adaptive Context-Aware Multi-Modal Network (ACMNet), achieves the state-of-the-art performance on two benchmarks.
  arXiv  Detail & Related papers  (2020-08-25T06:00:06Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.