arXiv:2603.25706v1 Announce Type: new
Abstract: Recent unified models have made unprecedented progress in both understanding and generation. However, while most of them accept multi-modal inputs, they typically produce only single-modality outputs. Th…
arXiv:2603.24709v1 Announce Type: cross
Abstract: Multi-step tool orchestration, where LLMs must invoke multiple dependent APIs in the correct order while propagating intermediate outputs, remains challenging. State-of-the-art models frequently fail o…
arXiv:2601.08881v2 Announce Type: replace
Abstract: Unified image generation and editing models suffer from severe task interference in dense diffusion transformers architectures, where a shared parameter space must compromise between conflicting obje…
arXiv:2603.24755v1 Announce Type: cross
Abstract: Software development is iterative, yet agentic coding benchmarks overwhelmingly evaluate single-shot solutions against complete specifications. Code can pass the test suite but become progressively har…
arXiv:2603.24649v1 Announce Type: new
Abstract: Currently, evaluating vision-language models (VLMs) in medical imaging tasks oversimplifies clinical reality by relying on pre-selected 2D images that demand significant manual labor to curate. This setu…
arXiv:2603.25140v1 Announce Type: new
Abstract: Multimodal deepfakes can exhibit subtle visual artifacts and cross-modal inconsistencies, which remain challenging to detect, especially when detectors are trained primarily on curated synthetic forgerie…
The European Parliament’s Thursday vote to delay parts of the EU AI Act adds more uncertainty to the already chaotic AI compliance universe. But analysts say that CIOs must proceed as though the compliance rules are in effect.
…
Apple’s MacBook Neo (reviewed here) challenges what we expect from budget laptops. Accompanied by shrewd enterprise-focused moves, the new model gives Apple a chance to convert hitherto resistant IT purchasers to adopt its platfo…
Intel’s Core Ultra Series 3 processors with Intel vPro, built for business PCs, are off to a fast start, already powering more than 125 designs including newly-announced systems from Dell and HP, the company said.
Unveiled thi…
Vision-language models (VLMs) use images and text to plan robot actions, but they still struggle to decide what actions to take and where to take them. Most systems split these decisions into two steps: a VLM generates a plan in natural language, and a separate model translates it into executable actions. This approach often breaks […]
The post GroundedPlanBench: Spatially grounded long-horizon task planning for robot manipulation appeared first on Microsoft Research.