Tim

Visual Prompt Engineering: How VIPE Improves Physical Reasoning on Abstract Images

by

Under DeepMind’s test conditions, abstract sketches may be one factor affecting a model’s physical-reasoning accuracy. VIPE automatically converts input images into a photorealistic style, raising Veo 3.1 accuracy on VPCT from 41.3% to 59.3%, while Omni Flash rises from 56.3% to 67.5%. Within the tested scope, this points to a direction: before spending heavily on more compute, first clarify what the model is actually seeing.

視覺提示工程：VIPE 如何改善抽象圖上的物理推理

by

在 DeepMind 的測試條件下，抽象草圖可能是影響模型物理推理準確率的因素之一。VIPE 透過自動將輸入圖像轉為寫實風格，讓 Veo 3.1 準確率從 41.3% 躍升至 59.3%。這在受測範圍內指出了一個方向：在花大錢增加算力之前，先搞清楚模型到底看到了什麼。

GMOS: The Trade-offs of Anchoring Moving Object Segmentation with 3D Geometry

by

Traditional moving object segmentation (MOS) relies on 2D aids such as optical flow, without 3D geometric context. GMOS anchors motion directly in RGB video across 3D space and time, introduces the frame-level MOS-I protocol, and builds the GMOS-2K dataset. The paper reports leading results across its evaluated MOS, MOS-I, and UVOS benchmarks, while running about three times faster than earlier multi-object MOS methods under the Tab. 1 evaluation setup.

Decoupled DiLoCo: How Decoupling Improves Resilience in Large-Scale Distributed Pre-training

by

Large language model pre-training can pause when hardware latency or node outages propagate through strict synchronization points. Decoupled DiLoCo uses independent Learners and asynchronous integration, allowing training to continue in simulated environments with millions of chips. Under the paper’s specified fault model, it reports zero global downtime and sustains 88% Goodput. The result is a more resilient architectural option for distributed training.

長程決策中的成對偏好強化學習：Markov Decision Contest 的機制與邊界

by

長程決策中，標量獎勵難以精確定義，而成對偏好方法在理論保證上存在缺口。研究提出的 Markov Decision Contest 模型將求解複雜度證明為 P 類，並提出 HPI 與 HPI-Clip 算法。實測顯示近似算法在多個高維任務中學習效率優於 SPPO，為長程成對偏好決策提供新的邊界與取捨。

GMOS：以 3D 幾何錨定動態物件分割的取捨

by

傳統動態物件分割（MOS）依賴光流等 2D 輔助，缺乏 3D 幾何資訊。GMOS 直接在 RGB 視頻上錨定 3D 時空，提出 MOS-I 逐幀瞬時運動評估協議並建立 GMOS-2K 資料集；論文報告其在所測 MOS、MOS-I 與 UVOS 基準中達到領先結果，並在 Tab. 1 的設定下比先前多物件 MOS 方法約快三倍。

Pairwise Preference Reinforcement Learning in Long-Horizon Decisions: The Mechanisms and Boundaries of Markov Decision Contest

by

In long-horizon decision-making, scalar rewards are difficult to define precisely, while pairwise preference methods have gaps in their theoretical guarantees. Markov Decision Contest proves that its exact solution is in P and introduces the HPI and HPI-Clip algorithms. Experiments across seven high-dimensional control tasks show that the approximate HPI-Clip method learns more efficiently than SPPO. The result defines a new set of boundaries and trade-offs for long-horizon decisions based on pairwise preferences.

When Does Stability Beat Plasticity in Gradually Non-Stationary Environments?

by

In gradually non-stationary environments, the main cause of reinforcement learning performance decline is instability caused by over-adaptation. This study finds that applying synaptic consolidation to multi-timescale Successor Features (SFs) can improve performance, but at the cost of computational latency and dependence on SGD optimizers.