SAS: Simulated Attention Score

The attention mechanism is a core component of the Transformer architecture. Various methods have been developed to compute attention scores, including multi-head attention (MHA), multi-query attention, group-query attention and so on. We further analyze the MHA and observe that its performance improves as the number of attention heads increases, provided the hidden size per head remains sufficiently large. Therefore, increasing both the head count and hidden size per head with minimal parameter overhead can lead to significant performance gains at a low cost. Motivated by this insight, we introduce Simulated Attention Score (SAS), which maintains a compact model size while simulating a larger number of attention heads and hidden feature dimension per head. This is achieved by projecting a low-dimensional head representation into a higher-dimensional space, effectively increasing attention capacity without increasing parameter count. Beyond the head representations, we further extend the simulation approach to feature dimension of the key and query embeddings, enhancing expressiveness by mimicking the behavior of a larger model while preserving the original model size. To control the parameter cost, we also propose Parameter-Efficient Attention Aggregation (PEAA). Comprehensive experiments on a variety of datasets and tasks demonstrate the effectiveness of the proposed SAS method, achieving significant improvements over different attention variants.

Decoder-Hybrid-Decoder Architecture for Efficient Reasoning with Long Generation

Recent advances in language modeling have demonstrated the effectiveness of State Space Models (SSMs) for efficient sequence modeling. While hybrid architectures such as Samba and the decoder-decoder architecture, YOCO, have shown promising performance gains over Transformers, prior works have not investigated the efficiency potential of representation sharing between SSM layers. In this paper, we introduce the Gated Memory Unit (GMU), a simple yet effective mechanism for efficient memory sharing across layers. We apply it to create SambaY, a decoder-hybrid-decoder architecture that incorporates GMUs in the cross-decoder to share memory readout states from a Samba-based self-decoder. SambaY significantly enhances decoding efficiency, preserves linear pre-filling time complexity, and boosts long-context performance, all while eliminating the need for explicit positional encoding. Through extensive scaling experiments, we demonstrate that our model exhibits a significantly lower irreducible loss compared to a strong YOCO baseline, indicating superior performance scalability under large-scale compute regimes. Our largest model enhanced with Differential Attention, Phi4-mini-Flash-Reasoning, achieves significantly better performance than Phi4-mini-Reasoning on reasoning tasks such as Math500, AIME24/25, and GPQA Diamond without any reinforcement learning, while delivering up to 10x higher decoding throughput on 2K-length prompts with 32K generation length under the vLLM inference framework. We release our training codebase on open-source data at https://github.com/microsoft/ArchScale.

SynapseRoute: An Auto-Route Switching Framework on Dual-State Large Language Model

With the widespread adoption of large language models (LLMs) in practical applications, selecting an appropriate model requires balancing not only performance but also operational cost. The emergence of reasoning-capable models has further widened the cost gap between "thinking" (high reasoning) and "non-thinking" (fast, low-cost) modes. In this work, we reveal that approximately 58% of medical questions can be accurately answered by the non-thinking mode alone, without requiring the high-cost reasoning process. This highlights a clear dichotomy in problem complexity and suggests that dynamically routing queries to the appropriate mode based on complexity could optimize accuracy, cost-efficiency, and overall user experience. Based on this, we further propose SynapseRoute, a machine learning-based dynamic routing framework that intelligently assigns input queries to either thinking or non-thinking modes. Experimental results on several medical datasets demonstrate that SynapseRoute not only improves overall accuracy (0.8390 vs. 0.8272) compared to the thinking mode alone but also reduces inference time by 36.8% and token consumption by 39.66%. Importantly, qualitative analysis indicates that over-reasoning on simpler queries can lead to unnecessary delays and even decreased accuracy, a pitfall avoided by our adaptive routing. Finally, this work further introduces the Accuracy-Inference-Token (AIT) index to comprehensively evaluate the trade-offs among accuracy, latency, and token cost.

SeqPO-SiMT: Sequential Policy Optimization for Simultaneous Machine Translation

We present Sequential Policy Optimization for Simultaneous Machine Translation (SeqPO-SiMT), a new policy optimization framework that defines the simultaneous machine translation (SiMT) task as a sequential decision making problem, incorporating a tailored reward to enhance translation quality while reducing latency. In contrast to popular Reinforcement Learning from Human Feedback (RLHF) methods, such as PPO and DPO, which are typically applied in single-step tasks, SeqPO-SiMT effectively tackles the multi-step SiMT task. This intuitive framework allows the SiMT LLMs to simulate and refine the SiMT process using a tailored reward. We conduct experiments on six datasets from diverse domains for En to Zh and Zh to En SiMT tasks, demonstrating that SeqPO-SiMT consistently achieves significantly higher translation quality with lower latency. In particular, SeqPO-SiMT outperforms the supervised fine-tuning (SFT) model by 1.13 points in COMET, while reducing the Average Lagging by 6.17 in the NEWSTEST2021 En to Zh dataset. While SiMT operates with far less context than offline translation, the SiMT results of SeqPO-SiMT on 7B LLM surprisingly rival the offline translation of high-performing LLMs, including Qwen-2.5-7B-Instruct and LLaMA-3-8B-Instruct.

Spatial RoboGrasp: Generalized Robotic Grasping Control Policy

Achieving generalizable and precise robotic manipulation across diverse environments remains a critical challenge, largely due to limitations in spatial perception. While prior imitation-learning approaches have made progress, their reliance on raw RGB inputs and handcrafted features often leads to overfitting and poor 3D reasoning under varied lighting, occlusion, and object conditions. In this paper, we propose a unified framework that couples robust multimodal perception with reliable grasp prediction. Our architecture fuses domain-randomized augmentation, monocular depth estimation, and a depth-aware 6-DoF Grasp Prompt into a single spatial representation for downstream action planning. Conditioned on this encoding and a high-level task prompt, our diffusion-based policy yields precise action sequences, achieving up to 40% improvement in grasp success and 45% higher task success rates under environmental variation. These results demonstrate that spatially grounded perception, paired with diffusion-based imitation learning, offers a scalable and robust solution for general-purpose robotic grasping.

Privacy Preserving Conversion Modeling in Data Clean Room

In the realm of online advertising, accurately predicting the conversion rate (CVR) is crucial for enhancing advertising efficiency and user satisfaction. This paper addresses the challenge of CVR prediction while adhering to user privacy preferences and advertiser requirements. Traditional methods face obstacles such as the reluctance of advertisers to share sensitive conversion data and the limitations of model training in secure environments like data clean rooms. We propose a novel model training framework that enables collaborative model training without sharing sample-level gradients with the advertising platform. Our approach introduces several innovative components: (1) utilizing batch-level aggregated gradients instead of sample-level gradients to minimize privacy risks; (2) applying adapter-based parameter-efficient fine-tuning and gradient compression to reduce communication costs; and (3) employing de-biasing techniques to train the model under label differential privacy, thereby maintaining accuracy despite privacy-enhanced label perturbations. Our experimental results, conducted on industrial datasets, demonstrate that our method achieves competitive ROCAUC performance while significantly decreasing communication overhead and complying with both advertiser privacy requirements and user privacy choices. This framework establishes a new standard for privacy-preserving, high-performance CVR prediction in the digital advertising landscape.

CoinRobot: Generalized End-to-end Robotic Learning for Physical Intelligence

Physical intelligence holds immense promise for advancing embodied intelligence, enabling robots to acquire complex behaviors from demonstrations. However, achieving generalization and transfer across diverse robotic platforms and environments requires careful design of model architectures, training strategies, and data diversity. Meanwhile existing systems often struggle with scalability, adaptability to heterogeneous hardware, and objective evaluation in real-world settings. We present a generalized end-to-end robotic learning framework designed to bridge this gap. Our framework introduces a unified architecture that supports cross-platform adaptability, enabling seamless deployment across industrial-grade robots, collaborative arms, and novel embodiments without task-specific modifications. By integrating multi-task learning with streamlined network designs, it achieves more robust performance than conventional approaches, while maintaining compatibility with varying sensor configurations and action spaces. We validate our framework through extensive experiments on seven manipulation tasks. Notably, Diffusion-based models trained in our framework demonstrated superior performance and generalizability compared to the LeRobot framework, achieving performance improvements across diverse robotic platforms and environmental conditions.

GRaD-Nav: Efficiently Learning Visual Drone Navigation with Gaussian Radiance Fields and Differentiable Dynamics

Autonomous visual navigation is an essential element in robot autonomy. Reinforcement learning (RL) offers a promising policy training paradigm. However existing RL methods suffer from high sample complexity, poor sim-to-real transfer, and limited runtime adaptability to navigation scenarios not seen during training. These problems are particularly challenging for drones, with complex nonlinear and unstable dynamics, and strong dynamic coupling between control and perception. In this paper, we propose a novel framework that integrates 3D Gaussian Splatting (3DGS) with differentiable deep reinforcement learning (DDRL) to train vision-based drone navigation policies. By leveraging high-fidelity 3D scene representations and differentiable simulation, our method improves sample efficiency and sim-to-real transfer. Additionally, we incorporate a Context-aided Estimator Network (CENet) to adapt to environmental variations at runtime. Moreover, by curriculum training in a mixture of different surrounding environments, we achieve in-task generalization, the ability to solve new instances of a task not seen during training. Drone hardware experiments demonstrate our method's high training efficiency compared to state-of-the-art RL methods, zero shot sim-to-real transfer for real robot deployment without fine tuning, and ability to adapt to new instances within the same task class (e.g. to fly through a gate at different locations with different distractors in the environment).

SIREN: Semantic, Initialization-Free Registration of Multi-Robot Gaussian Splatting Maps

We present SIREN for registration of multi-robot Gaussian Splatting (GSplat) maps, with zero access to camera poses, images, and inter-map transforms for initialization or fusion of local submaps. To realize these capabilities, SIREN harnesses the versatility and robustness of semantics in three critical ways to derive a rigorous registration pipeline for multi-robot GSplat maps. First, SIREN utilizes semantics to identify feature-rich regions of the local maps where the registration problem is better posed, eliminating the need for any initialization which is generally required in prior work. Second, SIREN identifies candidate correspondences between Gaussians in the local maps using robust semantic features, constituting the foundation for robust geometric optimization, coarsely aligning 3D Gaussian primitives extracted from the local maps. Third, this key step enables subsequent photometric refinement of the transformation between the submaps, where SIREN leverages novel-view synthesis in GSplat maps along with a semantics-based image filter to compute a high-accuracy non-rigid transformation for the generation of a high-fidelity fused map. We demonstrate the superior performance of SIREN compared to competing baselines across a range of real-world datasets, and in particular, across the most widely-used robot hardware platforms, including a manipulator, drone, and quadruped. In our experiments, SIREN achieves about 90x smaller rotation errors, 300x smaller translation errors, and 44x smaller scale errors in the most challenging scenes, where competing methods struggle. We will release the code and provide a link to the project page after the review process.

FAST-Splat: Fast, Ambiguity-Free Semantics Transfer in Gaussian Splatting

We present FAST-Splat for fast, ambiguity-free semantic Gaussian Splatting, which seeks to address the main limitations of existing semantic Gaussian Splatting methods, namely: slow training and rendering speeds; high memory usage; and ambiguous semantic object localization. In deriving FAST-Splat , we formulate open-vocabulary semantic Gaussian Splatting as the problem of extending closed-set semantic distillation to the open-set (open-vocabulary) setting, enabling FAST-Splat to provide precise semantic object localization results, even when prompted with ambiguous user-provided natural-language queries. Further, by exploiting the explicit form of the Gaussian Splatting scene representation to the fullest extent, FAST-Splat retains the remarkable training and rendering speeds of Gaussian Splatting. Specifically, while existing semantic Gaussian Splatting methods distill semantics into a separate neural field or utilize neural models for dimensionality reduction, FAST-Splat directly augments each Gaussian with specific semantic codes, preserving the training, rendering, and memory-usage advantages of Gaussian Splatting over neural field methods. These Gaussian-specific semantic codes, together with a hash-table, enable semantic similarity to be measured with open-vocabulary user prompts and further enable FAST-Splat to respond with unambiguous semantic object labels and 3D masks, unlike prior methods. In experiments, we demonstrate that FAST-Splat is 4x to 6x faster to train with a 13x faster data pre-processing step, achieves between 18x to 75x faster rendering speeds, and requires about 3x smaller GPU memory, compared to the best-competing semantic Gaussian Splatting methods. Further, FAST-Splat achieves relatively similar or better semantic segmentation performance compared to existing methods. After the review period, we will provide links to the project website and the codebase.