Entropy-Guided GRVQ for Ultra-Low Bitrate Neural Speech Codec

Neural audio codec (NAC) is essential for reconstructing high-quality speech signals and generating discrete representations for downstream speech language models. However, ensuring accurate semantic modeling while maintaining high-fidelity reconstruction under ultra-low bitrate constraints remains challenging. We propose an entropy-guided group residual vector quantization (EG-GRVQ) for an ultra-low bitrate neural speech codec, which retains a semantic branch for linguistic information and incorporates an entropy-guided grouping strategy in the acoustic branch. Assuming that channel activations follow approximately Gaussian statistics, the variance of each channel can serve as a principled proxy for its information content. Based on this assumption, we partition the encoder output such that each group carries an equal share of the total information. This balanced allocation improves codebook efficiency and reduces redundancy. Trained on LibriTTS and VCTK, our model shows improvements in perceptual quality and intelligibility metrics under ultra-low bitrate conditions, with a focus on codec-level fidelity for communication-oriented scenarios.

Geometry OR Tracker: Universal Geometric Operating Room Tracking

In operating rooms (OR), world-scale multi-view 3D tracking supports downstream applications such as surgeon behavior recognition, where physically meaningful quantities such as distances and motion statistics must be measured in meters. However, real clinical deployments rarely satisfy the geometric prerequisites for stable multi-view fusion and tracking: camera calibration and RGB-D registration are always unreliable, leading to cross-view geometric inconsistency that produces "ghosting" during fusion and degrades 3D trajectories in a shared OR coordinate frame. To address this, we introduce Geometry OR Tracker, a two-stage pipeline that first rectifies imprecise calibration into a scaleconsistent and geometrically consistent camera setup with a single global scale via a Multi-view Metric Geometry Rectification module, and then performs Occlusion-Robust 3D Point Tracking directly in the unified OR world frame. On the MM-OR benchmark, improved geometric consistency translates into tracking gains: our rectification front-end reduces cross-view depth disagreement by more than 30$\times$ compared to raw calibration. Ablation studies further demonstrate the relationship between calibration quality and tracking accuracy, showing that improved geometric consistency yields stronger world-frame tracking.

On the Mechanism and Dynamics of Modular Addition: Fourier Features, Lottery Ticket, and Grokking

We present a comprehensive analysis of how two-layer neural networks learn features to solve the modular addition task. Our work provides a full mechanistic interpretation of the learned model and a theoretical explanation of its training dynamics. While prior work has identified that individual neurons learn single-frequency Fourier features and phase alignment, it does not fully explain how these features combine into a global solution. We bridge this gap by formalizing a diversification condition that emerges during training when overparametrized, consisting of two parts: phase symmetry and frequency diversification. We prove that these properties allow the network to collectively approximate a flawed indicator function on the correct logic for the modular addition task. While individual neurons produce noisy signals, the phase symmetry enables a majority-voting scheme that cancels out noise, allowing the network to robustly identify the correct sum. Furthermore, we explain the emergence of these features under random initialization via a lottery ticket mechanism. Our gradient flow analysis proves that frequencies compete within each neuron, with the "winner" determined by its initial spectral magnitude and phase alignment. From a technical standpoint, we provide a rigorous characterization of the layer-wise phase coupling dynamics and formalize the competitive landscape using the ODE comparison lemma. Finally, we use these insights to demystify grokking, characterizing it as a three-stage process involving memorization followed by two generalization phases, driven by the competition between loss minimization and weight decay.

Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters

We introduce Step 3.5 Flash, a sparse Mixture-of-Experts (MoE) model that bridges frontier-level agentic intelligence and computational efficiency. We focus on what matters most when building agents: sharp reasoning and fast, reliable execution. Step 3.5 Flash pairs a 196B-parameter foundation with 11B active parameters for efficient inference. It is optimized with interleaved 3:1 sliding-window/full attention and Multi-Token Prediction (MTP-3) to reduce the latency and cost of multi-round agentic interactions. To reach frontier-level intelligence, we design a scalable reinforcement learning framework that combines verifiable signals with preference feedback, while remaining stable under large-scale off-policy training, enabling consistent self-improvement across mathematics, code, and tool use. Step 3.5 Flash demonstrates strong performance across agent, coding, and math tasks, achieving 85.4% on IMO-AnswerBench, 86.4% on LiveCodeBench-v6 (2024.08-2025.05), 88.2% on tau2-Bench, 69.0% on BrowseComp (with context management), and 51.0% on Terminal-Bench 2.0, comparable to frontier models such as GPT-5.2 xHigh and Gemini 3.0 Pro. By redefining the efficiency frontier, Step 3.5 Flash provides a high-density foundation for deploying sophisticated agents in real-world industrial environments.

Time2General: Learning Spatiotemporal Invariant Representations for Domain-Generalization Video Semantic Segmentation

Domain Generalized Video Semantic Segmentation (DGVSS) is trained on a single labeled driving domain and is directly deployed on unseen domains without target labels and test-time adaptation while maintaining temporally consistent predictions over video streams. In practice, both domain shift and temporal-sampling shift break correspondence-based propagation and fixed-stride temporal aggregation, causing severe frame-to-frame flicker even in label-stable regions. We propose Time2General, a DGVSS framework built on Stability Queries. Time2General introduces a Spatio-Temporal Memory Decoder that aggregates multi-frame context into a clip-level spatio-temporal memory and decodes temporally consistent per-frame masks without explicit correspondence propagation. To further suppress flicker and improve robustness to varying sampling rates, the Masked Temporal Consistency Loss is proposed to regularize temporal prediction discrepancies across different strides, and randomize training strides to expose the model to diverse temporal gaps. Extensive experiments on multiple driving benchmarks show that Time2General achieves a substantial improvement in cross-domain accuracy and temporal stability over prior DGSS and VSS baselines while running at up to 18 FPS. Code will be released after the review process.

Unlocking Out-of-Distribution Generalization in Transformers via Recursive Latent Space Reasoning

Systematic, compositional generalization beyond the training distribution remains a core challenge in machine learning -- and a critical bottleneck for the emergent reasoning abilities of modern language models. This work investigates out-of-distribution (OOD) generalization in Transformer networks using a GSM8K-style modular arithmetic on computational graphs task as a testbed. We introduce and explore a set of four architectural mechanisms aimed at enhancing OOD generalization: (i) input-adaptive recurrence; (ii) algorithmic supervision; (iii) anchored latent representations via a discrete bottleneck; and (iv) an explicit error-correction mechanism. Collectively, these mechanisms yield an architectural approach for native and scalable latent space reasoning in Transformer networks with robust algorithmic generalization capabilities. We complement these empirical results with a detailed mechanistic interpretability analysis that reveals how these mechanisms give rise to robust OOD generalization abilities.

SGS-3D: High-Fidelity 3D Instance Segmentation via Reliable Semantic Mask Splitting and Growing

Accurate 3D instance segmentation is crucial for high-quality scene understanding in the 3D vision domain. However, 3D instance segmentation based on 2D-to-3D lifting approaches struggle to produce precise instance-level segmentation, due to accumulated errors introduced during the lifting process from ambiguous semantic guidance and insufficient depth constraints. To tackle these challenges, we propose splitting and growing reliable semantic mask for high-fidelity 3D instance segmentation (SGS-3D), a novel "split-then-grow" framework that first purifies and splits ambiguous lifted masks using geometric primitives, and then grows them into complete instances within the scene. Unlike existing approaches that directly rely on raw lifted masks and sacrifice segmentation accuracy, SGS-3D serves as a training-free refinement method that jointly fuses semantic and geometric information, enabling effective cooperation between the two levels of representation. Specifically, for semantic guidance, we introduce a mask filtering strategy that leverages the co-occurrence of 3D geometry primitives to identify and remove ambiguous masks, thereby ensuring more reliable semantic consistency with the 3D object instances. For the geometric refinement, we construct fine-grained object instances by exploiting both spatial continuity and high-level features, particularly in the case of semantic ambiguity between distinct objects. Experimental results on ScanNet200, ScanNet++, and KITTI-360 demonstrate that SGS-3D substantially improves segmentation accuracy and robustness against inaccurate masks from pre-trained models, yielding high-fidelity object instances while maintaining strong generalization across diverse indoor and outdoor environments. Code is available in the supplementary materials.

EGS-SLAM: RGB-D Gaussian Splatting SLAM with Events

Gaussian Splatting SLAM (GS-SLAM) offers a notable improvement over traditional SLAM methods, enabling photorealistic 3D reconstruction that conventional approaches often struggle to achieve. However, existing GS-SLAM systems perform poorly under persistent and severe motion blur commonly encountered in real-world scenarios, leading to significantly degraded tracking accuracy and compromised 3D reconstruction quality. To address this limitation, we propose EGS-SLAM, a novel GS-SLAM framework that fuses event data with RGB-D inputs to simultaneously reduce motion blur in images and compensate for the sparse and discrete nature of event streams, enabling robust tracking and high-fidelity 3D Gaussian Splatting reconstruction. Specifically, our system explicitly models the camera's continuous trajectory during exposure, supporting event- and blur-aware tracking and mapping on a unified 3D Gaussian Splatting scene. Furthermore, we introduce a learnable camera response function to align the dynamic ranges of events and images, along with a no-event loss to suppress ringing artifacts during reconstruction. We validate our approach on a new dataset comprising synthetic and real-world sequences with significant motion blur. Extensive experimental results demonstrate that EGS-SLAM consistently outperforms existing GS-SLAM systems in both trajectory accuracy and photorealistic 3D Gaussian Splatting reconstruction. The source code will be available at https://github.com/Chensiyu00/EGS-SLAM.

ICM-Fusion: In-Context Meta-Optimized LoRA Fusion for Multi-Task Adaptation

Enabling multi-task adaptation in pre-trained Low-Rank Adaptation (LoRA) models is crucial for enhancing their generalization capabilities. Most existing pre-trained LoRA fusion methods decompose weight matrices, sharing similar parameters while merging divergent ones. However, this paradigm inevitably induces inter-weight conflicts and leads to catastrophic domain forgetting. While incremental learning enables adaptation to multiple tasks, it struggles to achieve generalization in few-shot scenarios. Consequently, when the weight data follows a long-tailed distribution, it can lead to forgetting in the fused weights. To address this issue, we propose In-Context Meta LoRA Fusion (ICM-Fusion), a novel framework that synergizes meta-learning with in-context adaptation. The key innovation lies in our task vector arithmetic, which dynamically balances conflicting optimization directions across domains through learned manifold projections. ICM-Fusion obtains the optimal task vector orientation for the fused model in the latent space by adjusting the orientation of the task vectors. Subsequently, the fused LoRA is reconstructed by a self-designed Fusion VAE (F-VAE) to realize multi-task LoRA generation. We have conducted extensive experiments on visual and linguistic tasks, and the experimental results demonstrate that ICM-Fusion can be adapted to a wide range of architectural models and applied to various tasks. Compared to the current pre-trained LoRA fusion method, ICM-Fusion fused LoRA can significantly reduce the multi-tasking loss and can even achieve task enhancement in few-shot scenarios.

Step-Audio 2 Technical Report

This paper presents Step-Audio 2, an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation. By integrating a latent audio encoder and reasoning-centric reinforcement learning (RL), Step-Audio 2 achieves promising performance in automatic speech recognition (ASR) and audio understanding. To facilitate genuine end-to-end speech conversation, Step-Audio 2 incorporates the generation of discrete audio tokens into language modeling, significantly enhancing its responsiveness to paralinguistic information such as speaking styles and emotions. To effectively leverage the rich textual and acoustic knowledge in real-world data, Step-Audio 2 integrates retrieval-augmented generation (RAG) and is able to call external tools such as web search to mitigate hallucination and audio search to switch timbres. Trained on millions of hours of speech and audio data, Step-Audio 2 delivers intelligence and expressiveness across diverse conversational scenarios. Evaluation results demonstrate that Step-Audio 2 achieves state-of-the-art performance on various audio understanding and conversational benchmarks compared to other open-source and commercial solutions. Please visit https://github.com/stepfun-ai/Step-Audio2 for more information.