arXiv subjects

Sirnam Swetha

Publications and source records attributed to Sirnam Swetha.

10 recordsLinked to original sources

arXiv2026

The Telephone Game: Evaluating Semantic Drift in Unified Models

Unified models (UMs) combine visual understanding (I2T) and generation (T2I) in a single framework. We focus on T2I and I2T, where cross-consistency---what a model understands, it should be able to generate---is a promise of unification and a necessity when composing both capabilities. Yet, existing benchmarks evaluate them in isolation: FID/GenEval for T2I; MME/MMBench for I2T. We show this gap is consequential: models scoring competitively on these benchmarks can fail severely when understanding and generation are composed, losing entities, attributes, spatial relations, and counts, resulting in semantic drift. To quantify drift, we introduce the Semantic Drift Protocol (SDP), inspired by the Telephone Game: starting from a caption or image, we alternate I2T and T2I over multiple generations and measure semantic preservation. We propose Mean Cumulative Drift (MCD), an embedding-based measure of content retention across three representation spaces, and Multi-Generation GenEval (MGG), extending GenEval's object-level compliance scoring across generations. To stress-test models beyond COCO-style data, we create a benchmark of 400 image-text pairs sampled from NoCaps and DOCCI, emphasizing novel objects and fine-grained descriptions. Applying SDP to seven models reveals that drift varies dramatically and is not predicted by single-pass scores: BAGEL retains high semantic fidelity over multiple generations, while VILA-U and Janus variants collapse within five generations, despite comparable isolated metrics. We identify six recurring failure modes and find degradation is typically catastrophic rather than gradual: once a critical error occurs, subsequent generations compound it. SDP exposes failure modes that single-pass benchmarks miss, enabling a more faithful assessment of unified model reliability. Code and benchmark: https://github.com/mollahsabbir/telephone-game-semantic-drift

cs.CV↗

arXiv2026

VRR-QA: Visual Relational Reasoning in Videos Beyond Explicit Cues

Video Question Answering (VideoQA) has made significant strides by leveraging multimodal learning to align visual and textual modalities. However, current benchmarks overwhelmingly focus on questions answerable through explicit visual content - actions, objects, and events - directly observable within individual frames or short clips. To truly understand videos as humans do, models must go beyond what is directly shown, inferring hidden relationships and contextual cues that are only implied across frames. Current benchmarks fail to capture this essential aspect of video understanding. To address this gap, we introduce VRR-QA, a benchmark for Visual Relational Reasoning Beyond Explicit Cues. We curate our benchmark from creative and cinematic videos such as movies, that deliberately employ storytelling techniques which omit direct depictions of certain events or relations, requiring viewers to infer them. VRR-QA comprises 1K meticulously expert-annotated QA pairs drawn from 1K creative video clips covering 15 genres across 7 decades of content, from both live-action and animated titles. Our extensive evaluations on 14 leading VideoQA models reveals consistent and significant performance degradation, underscoring their reliance on surface-level visual cues and highlighting the difficulty of implicit reasoning. Even the best model substantially underperforms human baselines with only 64% accuracy. Performance variations across models further illustrate the complexity and diversity of the challenges presented by VRR-QA. By releasing both dataset and data collection framework, VRR-QA establishes a rigorous, diverse, and reproducible testbed for advancing VideoQA: https://swetha5.github.io/ImplicitQA/.

cs.CV↗

arXiv2026

TIGeR: A Unified Framework for Time, Images and Geo-location Retrieval

Many real-world applications in digital forensics, urban monitoring, and environmental analysis require jointly reasoning about visual appearance, location, and time. Beyond standard geo-localization and time-of-capture prediction, these applications increasingly demand more complex capabilities, such as retrieving an image captured at the same location as a query image but at a specified target time. We formalize this problem as Geo-Time Aware Image Retrieval and propose TIGeR, a unified framework for Time, Images and Geo-location Retrieval. TIGeR supports flexible input configurations (single-modality and multi-modality queries) and uses the same representation to perform (i) geo-localization, (ii) time-of-capture prediction, and (iii) geo-time-aware retrieval. By preserving the underlying location identity despite large appearance changes, TIGeR enables retrieval based on where and when a scene was captured, rather than purely on visual similarity. To support this task, we design a multistage data curation pipeline and propose a new diverse dataset of 4.5M paired image-location-time triplets for training and 86k high-quality triplets for evaluation. Extensive experiments show that TIGeR consistently outperforms strong baselines and state-of-the-art methods by up to 16% on time-of-year, 8% time-of-day prediction, and 14% in geo-time aware retrieval recall, highlighting the benefits of unified geo-temporal modeling.

cs.CV↗

arXiv2026

BBQ-V: Benchmarking Visual Stereotype Bias in Large Multimodal Models

Stereotype biases in Large Multimodal Models (LMMs) perpetuate harmful societal prejudices, undermining the fairness and equity of AI applications. As LMMs grow increasingly influential, addressing and mitigating inherent biases related to stereotypes, harmful generations, and ambiguous assumptions in real-world scenarios has become essential. However, existing datasets evaluating stereotype biases in LMMs often lack diversity, rely on synthetic images, and often have single-actor images, leaving a gap in bias evaluation for real-world visual contexts. To address the gap in bias evaluation using real images, we introduce the BBQ-Vision (BBQ-V), the most comprehensive framework for assessing stereotype biases across nine diverse categories and 50 sub-categories with real and multi-actor images. BBQ-V benchmark contains 14,144 image-question pairs and rigorously evaluates LMMs through carefully curated, visually grounded scenarios, challenging them to reason accurately about visual stereotypes. It offers a robust evaluation framework featuring real-world visual samples, image variations, and open-ended question formats. BBQ-V enables a precise and nuanced assessment of a model's reasoning capabilities across varying levels of difficulty. Through rigorous testing of 19 state-of-the-art open-source (general-purpose and reasoning) and closed-source LMMs, we highlight that these top-performing models are often biased on several social stereotypes, and demonstrate that the thinking models induce more bias in the reasoning chains. This benchmark represents a significant step toward fostering fairness in AI systems and reducing harmful biases, laying the groundwork for more equitable and socially responsible LMMs. Our dataset and evaluation code are publicly available.

cs.CV↗

arXiv2025

StretchySnake: Flexible SSM Training Unlocks Action Recognition Across Spatio-Temporal Scales

State space models (SSMs) have emerged as a competitive alternative to transformers in various tasks. Their linear complexity and hidden-state recurrence make them particularly attractive for modeling long sequences, whereas attention becomes quadratically expensive. However, current training methods for video understanding are tailored towards transformers and fail to fully leverage the unique attributes of SSMs. For example, video models are often trained at a fixed resolution and video length to balance the quadratic scaling of attention cost against performance. Consequently, these models suffer from degraded performance when evaluated on videos with spatial and temporal resolutions unseen during training; a property we call spatio-temporal inflexibility. In the context of action recognition, this severely limits a model's ability to retain performance across both short- and long-form videos. Therefore, we propose a flexible training method that leverages and improves the inherent adaptability of SSMs. Our method samples videos at varying temporal and spatial resolutions during training and dynamically interpolates model weights to accommodate any spatio-temporal scale. This instills our SSM, which we call StretchySnake, with spatio-temporal flexibility and enables it to seamlessly handle videos ranging from short, fine-grained clips to long, complex activities. We introduce and compare five different variants of flexible training, and identify the most effective strategy for video SSMs. On short-action (UCF-101, HMDB-51) and long-action (COIN, Breakfast) benchmarks, StretchySnake outperforms transformer and SSM baselines alike by up to 28%, with strong adaptability to fine-grained actions (SSV2, Diving-48). Therefore, our method provides a simple drop-in training recipe that makes video SSMs more robust, resolution-agnostic, and efficient across diverse action recognition scenarios.

cs.CV↗

arXiv2025

Safe-LLaVA: A Privacy-Preserving Vision-Language Dataset and Benchmark for Biometric Safety

Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities in vision-language tasks. However, these models often infer and reveal sensitive biometric attributes such as race, gender, age, body weight, and eye color; even when such information is not explicitly requested. This raises critical concerns, particularly in real-world applications and socially-sensitive domains. Despite increasing awareness, no publicly available dataset or benchmark exists to comprehensively evaluate or mitigate biometric leakage in MLLMs. To address this gap, we introduce PRISM (Privacy-aware Evaluation of Responses in Sensitive Modalities), a new benchmark designed to assess MLLMs on two fronts: (1) refuse biometric-related queries and (2) implicit biometric leakage in general responses while maintaining semantic faithfulness. Further, we conduct a detailed audit of the widely used LLaVA datasets and uncover extensive biometric leakage across pretraining and instruction data. To address this, we present Safe-LLaVA dataset, the first privacy-preserving MLLM training dataset constructed by systematically removing explicit and implicit biometric information from LLaVA dataset. Our evaluations on PRISM reveal biometric leakages across MLLMs for different attributes, highlighting the detailed privacy-violations. We also fine-tune a model on Safe-LLaVA dataset and show that it substantially reduces the biometric leakages. Together, Safe-LLaVA and PRISM set a new standard for privacy-aligned development and evaluation of MLLMs.

cs.CV↗

arXiv2025

GT-Loc: Unifying When and Where in Images Through a Joint Embedding Space

Timestamp prediction aims to determine when an image was captured using only visual information, supporting applications such as metadata correction, retrieval, and digital forensics. In outdoor scenarios, hourly estimates rely on cues like brightness, hue, and shadow positioning, while seasonal changes and weather inform date estimation. However, these visual cues significantly depend on geographic context, closely linking timestamp prediction to geo-localization. To address this interdependence, we introduce GT-Loc, a novel retrieval-based method that jointly predicts the capture time (hour and month) and geo-location (GPS coordinates) of an image. Our approach employs separate encoders for images, time, and location, aligning their embeddings within a shared high-dimensional feature space. Recognizing the cyclical nature of time, instead of conventional contrastive learning with hard positives and negatives, we propose a temporal metric-learning objective providing soft targets by modeling pairwise time differences over a cyclical toroidal surface. We present new benchmarks demonstrating that our joint optimization surpasses previous time prediction methods, even those using the ground-truth geo-location as an input during inference. Additionally, our approach achieves competitive results on standard geo-localization tasks, and the unified embedding space facilitates compositional and text-based image retrieval.

cs.CV↗

arXiv2025

TimeLogic: A Temporal Logic Benchmark for Video QA

Temporal logical understanding, a core facet of human cognition, plays a pivotal role in capturing complex sequential events and their temporal relationships within videos. This capability is particularly crucial in tasks like Video Question Answering (VideoQA), where the goal is to process visual data over time together with textual data to provide coherent answers. However, current VideoQA benchmarks devote little focus to evaluating this critical skill due to the challenge of annotating temporal logic. Despite the advancement of vision-language models, assessing their temporal logical reasoning powers remains a challenge, primarily due to the lack QA pairs that demand formal, complex temporal reasoning. To bridge this gap, we introduce the TimeLogic QA (TLQA) framework to automatically generate the QA pairs, specifically designed to evaluate the temporal logical understanding. To this end, TLQA leverages temporal annotations from existing video datasets together with temporal operators derived from logic theory to construct questions that test understanding of event sequences and their temporal relationships. TLQA framework is generic and scalable, capable of leveraging both, existing video action datasets with temporal action segmentation annotations, or video datasets with temporal scene graph annotations, to automatically generate temporal logical questions. We leverage 4 datasets, STAR, Breakfast, AGQA, and CrossTask, and generate two VideoQA dataset variants - small (TLQA-S) and large (TLQA-L) - containing 2k and 10k QA pairs for each category, resulting in 32k and 160k total pairs per dataset. We undertake a comprehensive evaluation of leading-edge VideoQA models, employing the TLQA to benchmark their temporal logical understanding capabilities. We assess the VideoQA model's temporal reasoning performance on 16 categories of temporal logic with varying temporal complexity.

cs.CV↗

arXiv2024

X-Former: Unifying Contrastive and Reconstruction Learning for MLLMs

Recent advancements in Multimodal Large Language Models (MLLMs) have revolutionized the field of vision-language understanding by integrating visual perception capabilities into Large Language Models (LLMs). The prevailing trend in this field involves the utilization of a vision encoder derived from vision-language contrastive learning (CL), showing expertise in capturing overall representations while facing difficulties in capturing detailed local patterns. In this work, we focus on enhancing the visual representations for MLLMs by combining high-frequency and detailed visual representations, obtained through masked image modeling (MIM), with semantically-enriched low-frequency representations captured by CL. To achieve this goal, we introduce X-Former which is a lightweight transformer module designed to exploit the complementary strengths of CL and MIM through an innovative interaction mechanism. Specifically, X-Former first bootstraps vision-language representation learning and multimodal-to-multimodal generative learning from two frozen vision encoders, i.e., CLIP-ViT (CL-based) and MAE-ViT (MIM-based). It further bootstraps vision-to-language generative learning from a frozen LLM to ensure visual features from X-Former can be interpreted by the LLM. To demonstrate the effectiveness of our approach, we assess its performance on tasks demanding detailed visual understanding. Extensive evaluations indicate that X-Former excels in visual reasoning tasks involving both structural and semantic categories in the GQA dataset. Assessment on fine-grained visual perception benchmark further confirms its superior capabilities in visual understanding.

cs.CV↗

arXiv2021

Unsupervised Discriminative Embedding for Sub-Action Learning in Complex Activities

Action recognition and detection in the context of long untrimmed video sequences has seen an increased attention from the research community. However, annotation of complex activities is usually time consuming and challenging in practice. Therefore, recent works started to tackle the problem of unsupervised learning of sub-actions in complex activities. This paper proposes a novel approach for unsupervised sub-action learning in complex activities. The proposed method maps both visual and temporal representations to a latent space where the sub-actions are learnt discriminatively in an end-to-end fashion. To this end, we propose to learn sub-actions as latent concepts and a novel discriminative latent concept learning (DLCL) module aids in learning sub-actions. The proposed DLCL module lends on the idea of latent concepts to learn compact representations in the latent embedding space in an unsupervised way. The result is a set of latent vectors that can be interpreted as cluster centers in the embedding space. The latent space itself is formed by a joint visual and temporal embedding capturing the visual similarity and temporal ordering of the data. Our joint learning with discriminative latent concept module is novel which eliminates the need for explicit clustering. We validate our approach on three benchmark datasets and show that the proposed combination of visual-temporal embedding and discriminative latent concepts allow to learn robust action representations in an unsupervised setting.

cs.CV↗