AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers propose Visual-Noise Guided In-Context Distillation (VGID), a novel framework for removing sensitive knowledge from multimodal large language models without full retraining. The method combines visual perturbation with textual in-context unlearning to achieve parameter-level knowledge removal while maintaining model performance, addressing critical privacy and safety concerns in MLLMs.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers introduced WebIGBench, the first benchmark for evaluating multimodal LLMs on code generation for interactive webpages, addressing a critical gap in existing evaluation frameworks that only assess static pages. The benchmark includes 103 real-world webpages with 871 distinct interactive actions and proposes novel automated assessment methods to measure interaction consistency beyond visual fidelity.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers identify and address Perceptual Judgment Bias in multimodal large language models used as automated evaluators, where these models favor plausible narratives over visually accurate answers when text and images conflict. The team develops a training framework using perceptually perturbed datasets and reward modeling that improves MLLM judges' visual grounding and evaluation consistency across benchmarks.
AIBullisharXiv – CS AI · Jun 26/10
🧠Researchers introduce DenseMLLM, a multimodal large language model that performs fine-grained dense prediction tasks like semantic segmentation and depth estimation without requiring task-specific decoders. The minimalist approach achieves competitive performance while maintaining the generalist design philosophy of standard MLLMs, potentially simplifying model architecture and increasing practical applicability.
AIBullisharXiv – CS AI · Jun 16/10
🧠Researchers introduce SCALE, a self-improving web agent framework that uses adversarial roles and cognitive-aware exploration to autonomously adapt to complex web environments without relying on handcrafted pipelines or expensive expert data. The framework includes SCALE-Hop, a graph exploration strategy, and SCALE-20k, a 20,000-sample dataset from 19 real-world websites that demonstrates improved performance across multiple multimodal large language models.
AIBullisharXiv – CS AI · Jun 16/10
🧠Researchers introduce MechVQA, the first comprehensive dataset for evaluating multimodal large language models (MLLMs) on mechanical drawing understanding, containing 3.3k annotated drawings with 21k question-answer pairs across three capability levels. They develop MechVL, a domain-specialized model that outperforms existing baselines by 7.57 percentage points, establishing a foundation for deploying AI in mechanical design and engineering inspection workflows.
AINeutralarXiv – CS AI · Jun 16/10
🧠Researchers introduce ERGeoBench, a comprehensive benchmark for evaluating multimodal large language models (MLLMs) on embodied geo-localization tasks using 2,207 street-view panoramas across three progressive difficulty settings. The evaluation reveals that current leading models can understand high-level geographic semantics but struggle with fine-grained perception, metric localization, and spatial consistency, highlighting that accurate geo-localization requires integrated perception and reasoning rather than isolated visual recognition.
AINeutralarXiv – CS AI · May 296/10
🧠Researchers propose LDKE, a new framework for editing knowledge in Multimodal Large Language Models that addresses two critical failure modes: causal misalignment (edits confined to specific samples) and feature entanglement (unintended alterations to related information). The method uses localized layer identification and input disentanglement to enable precise, generalized edits while preserving unrelated knowledge.
AINeutralarXiv – CS AI · May 296/10
🧠Researchers introduce PlanAhead, a framework that systematically evaluates how different natural language plan representations affect LLM-based web agent performance across multiple AI models. The study finds that both the plan formulation method and underlying LLM significantly impact agent robustness, with implications for improving autonomous AI systems that interact with web interfaces.
🏢 OpenAI
AINeutralarXiv – CS AI · May 296/10
🧠Researchers present Empathic Prompting, a framework that integrates facial expression recognition into multimodal LLM conversations to capture and embed users' emotional cues as contextual signals. The system operates unobtrusively through a locally deployed DeepSeek instance and demonstrates coherent integration of non-verbal input in a preliminary evaluation (N=5), with potential applications in healthcare and education.
AIBullisharXiv – CS AI · May 286/10
🧠Researchers introduce MOV-Bench, a benchmark for evaluating multi-hop audio-visual reasoning in large language models, and propose AOP-Agent, an agentic framework that enables open-source multimodal LLMs to perform active perception across temporally dispersed audio and visual evidence without additional training.
AIBullisharXiv – CS AI · May 276/10
🧠Researchers introduce POLAR, a memory-augmented framework that enables multimodal AI agents to personalize their behavior based on accumulated long-term user interactions. The system organizes past interactions into semantic and episodic memory, allowing embodied agents to interpret implicit user requests and improve task execution performance across multiple interaction cycles.
AINeutralarXiv – CS AI · May 276/10
🧠Researchers introduced PhyWorldBench, a comprehensive benchmark that evaluates text-to-video generation models on their ability to simulate real-world physics accurately. Testing 12 state-of-the-art models across 1,050 prompts, the study reveals significant gaps in how current AI video generators handle physical phenomena, from basic object motion to complex interactions, while also introducing novel evaluation methods using multimodal language models.
AINeutralarXiv – CS AI · May 126/10
🧠Researchers propose SFFL, a framework that mitigates cross-modal interference in audio-visual language models by enforcing separate reasoning chains for each modality before fusion. The approach uses modality-preference labels and reinforcement learning to reduce hallucinations and achieves 5-11% performance improvements on benchmarks.
AINeutralarXiv – CS AI · May 126/10
🧠Researchers introduce STEMO-Bench, a benchmark for evaluating video understanding in multimodal large language models (MLLMs), and propose STEMO-Track, a framework that reduces hallucinations by explicitly tracking object identities and states across time. The work addresses a critical limitation in current video AI systems: their inability to persistently monitor objects and temporal relationships in dynamic scenes.
AINeutralarXiv – CS AI · May 116/10
🧠Researchers introduce EgoPro-Bench, a comprehensive benchmark dataset with over 14,000 egocentric videos designed to train and evaluate proactive AI assistants that can understand user intent and interact at optimal moments. The work addresses limitations in existing multimodal large language models by enabling personalized, timing-aware interactions rather than purely reactive responses.
AINeutralarXiv – CS AI · May 96/10
🧠Researchers developed a causal probing framework to decode how Multimodal Large Language Models internally represent visual concepts, revealing that entities are encoded in localized regions while abstract concepts distribute globally across networks. The findings expose mechanistic drivers of scaling laws and uncover a disconnect between visual perception and reasoning capabilities in MLLMs.
AINeutralarXiv – CS AI · May 96/10
🧠Researchers present a novel machine unlearning approach for Multimodal Large Language Models that selectively removes target visual knowledge while preserving non-target information across both visual and textual modalities. The method uses contrastive visual forgetting and null space constraints to balance effective forgetting with knowledge retention, extending applicability to continual unlearning scenarios.
AINeutralarXiv – CS AI · May 96/10
🧠Researchers introduce CrossCult-KIBench, a benchmark dataset for evaluating how multimodal large language models (MLLMs) handle cross-cultural knowledge insertion across English, Chinese, and Arabic contexts. The work reveals that current AI models struggle to adapt to specific cultural contexts without degrading performance in other cultures, establishing a new research direction for culturally-aware AI systems.
AIBullisharXiv – CS AI · May 76/10
🧠Pro²Assist is a step-aware AI assistant that uses augmented reality glasses and multimodal perception to provide real-time, proactive guidance for multi-step procedural tasks. The system tracks user progress continuously and demonstrates 21% higher accuracy in action understanding and 2.29x better timing accuracy compared to existing baselines, with 90% user approval in testing.
AINeutralarXiv – CS AI · May 16/10
🧠Researchers introduced COHERENCE, a new benchmark for evaluating Multimodal Large Language Models (MLLMs) on their ability to understand fine-grained image-text alignment in interleaved contexts—such as documents with mixed text and images. The benchmark contains 6,161 high-quality questions across four domains and includes error analysis to identify specific capability gaps in current models.
AINeutralarXiv – CS AI · Apr 206/10
🧠Researchers introduced 'Mind's Eye,' a benchmark that tests multimodal large language models (MLLMs) on visual reasoning tasks inspired by human intelligence tests. The evaluation reveals a significant gap between human performance (80% accuracy) and leading MLLMs (below 50%), exposing limitations in visuospatial reasoning, visual attention, and conceptual abstraction.
AIBullisharXiv – CS AI · Apr 156/10
🧠Researchers introduce CLASP, a token reduction framework that optimizes Multimodal Large Language Models by intelligently pruning visual tokens through class-adaptive layer fusion and dual-stage pruning. The approach addresses computational inefficiency in MLLMs while maintaining performance across diverse benchmarks and architectures.
AINeutralarXiv – CS AI · Apr 146/10
🧠Researchers introduce Diffusion-CAM, a novel interpretability method designed specifically for diffusion-based Multimodal Large Language Models (dMLLMs). Unlike existing visualization techniques optimized for sequential models, this approach accounts for the parallel denoising process inherent to diffusion architectures, achieving superior localization accuracy and visual fidelity in model explanations.
AIBullisharXiv – CS AI · Apr 146/10
🧠Researchers introduce MCERF, a multimodal retrieval framework that combines vision-language models with LLM reasoning to improve question-answering from engineering documents. The system achieves a 41.1% relative accuracy improvement over baseline RAG systems by handling complex multimodal content like tables, diagrams, and dense technical text through adaptive routing and hybrid retrieval strategies.
