Walking assistance in extreme or complex environments remains a significant challenge for people with blindness or low vision (BLV), largely due to the lack of a holistic scene understanding. Motivated by the real-world needs of the BLV community, we build mmWalk, a simulated multi-modal dataset that integrates multi-view sensor and accessibility-oriented features for outdoor safe navigation. Our dataset comprises 120manually controlled, scenario-categorized walking trajectories with 62k synchronized frames. It contains over 559k panoramic images across RGB, depth, and semantic modalities. Furthermore, to emphasize realworld relevance, each trajectory involves outdoor corner cases and accessibilityspecific landmarks for BLV users. Additionally, we generate mmWalkVQA, a VQA benchmark with over 69kvisual question-answer triplets across 9categories tailored for safe and informed walking assistance. We evaluate state-of-the-art Vision-Language Models (VLMs) using zero-and few-shot settings and found they struggle with our risk assessment and navigational tasks. We validate our mmWalk-finetuned model on real-world datasets and show the effectiveness of our dataset for advancing multi-modal walking assistance.

As Large Language Models (LLMs) become more capable of handling increasingly complex tasks, the evaluation set must keep pace with these advancements to ensure it remains sufficiently discriminative. Item Discrimination (ID) theory, which is widely used in educational assessment, measures the ability of individual test items to differentiate between high and low performers. Inspired by this theory, we propose an ID-induced prompt synthesis framework for evaluating LLMs so that the evaluation set continually updates and refines according to model abilities.

Video game fans and big, blockbuster releases have had an uneasy relationship in recent years. As so-called triple-A games get more expensive to make, the publishers behind them are accused of taking fewer risks and failing to try new things. But highly anticipated new release Crimson Desert asks a different question - what if a big-budget, graphically advanced game tried to do absolutely everything? The ambitious action-adventure's been compared to a buffet, presenting players with a smorgasbord of ideas, gameplay styles and quests to gorge on. While some have praised it as a feast, others have found it overstuffed, with some undercooked morsels behind the impressive presentation.

Large-scale pre-trained language models have achieved impressive results on a wide range of downstream tasks recently.

The agents often have difficulties in cooperating on common goals, dividing complex tasks, and planning through several stages to make progress.

Neural Information Processing Systems http://nips.cc/