Power Homotopy for Zeroth-Order Non-Convex Optimizations

We introduce GS-PowerHP, a novel zeroth-order method for non-convex optimization problems of the form $\max_{x \in \mathbb{R}^d} f(x)$. Our approach leverages two key components: a power-transformed Gaussian-smoothed surrogate $F_{N,σ}(μ) = \mathbb{E}_{x\sim\mathcal{N}(μ,σ^2 I_d)}[e^{N f(x)}]$ whose stationary points cluster near the global maximizer $x^*$ of $f$ for sufficiently large $N$, and an incrementally decaying $σ$ for enhanced data efficiency. Under mild assumptions, we prove convergence in expectation to a small neighborhood of $x^*$ with the iteration complexity of $O(d^2 \varepsilon^{-2})$. Empirical results show our approach consistently ranks among the top three across a suite of competing algorithms. Its robustness is underscored by the final experiment on a substantially high-dimensional problem ($d=150,528$), where it achieved first place on least-likely targeted black-box attacks against images from ImageNet, surpassing all competing methods.

Is It Truly Necessary to Process and Fit Minutes-Long Reference Videos for Personalized Talking Face Generation?

Talking Face Generation (TFG) aims to produce realistic and dynamic talking portraits, with broad applications in fields such as digital education, film and television production, e-commerce live streaming, and other related areas. Currently, TFG methods based on Neural Radiated Field (NeRF) or 3D Gaussian sputtering (3DGS) are received widespread attention. They learn and store personalized features from reference videos of each target individual to generate realistic speaking videos. To ensure models can capture sufficient 3D information and successfully learns the lip-audio mapping, previous studies usually require meticulous processing and fitting several minutes of reference video, which always takes hours. The computational burden of processing and fitting long reference videos severely limits the practical application value of these methods.However, is it really necessary to fit such minutes of reference video? Our exploratory case studies show that using some informative reference video segments of just a few seconds can achieve performance comparable to or even better than the full reference video. This indicates that video informative quality is much more important than its length. Inspired by this observation, we propose the ISExplore (short for Informative Segment Explore), a simple-yet-effective segment selection strategy that automatically identifies the informative 5-second reference video segment based on three key data quality dimensions: audio feature diversity, lip movement amplitude, and number of camera views. Extensive experiments demonstrate that our approach increases data processing and training speed by more than 5x for NeRF and 3DGS methods, while maintaining high-fidelity output. Project resources are available at xx.

LLM-as-a-Supervisor: Mistaken Therapeutic Behaviors Trigger Targeted Supervisory Feedback

Although large language models (LLMs) hold significant promise in psychotherapy, their direct application in patient-facing scenarios raises ethical and safety concerns. Therefore, this work shifts towards developing an LLM as a supervisor to train real therapists. In addition to the privacy of clinical therapist training data, a fundamental contradiction complicates the training of therapeutic behaviors: clear feedback standards are necessary to ensure a controlled training system, yet there is no absolute "gold standard" for appropriate therapeutic behaviors in practice. In contrast, many common therapeutic mistakes are universal and identifiable, making them effective triggers for targeted feedback that can serve as clearer evidence. Motivated by this, we create a novel therapist-training paradigm: (1) guidelines for mistaken behaviors and targeted correction strategies are first established as standards; (2) a human-in-the-loop dialogue-feedback dataset is then constructed, where a mistake-prone agent intentionally makes standard mistakes during interviews naturally, and a supervisor agent locates and identifies mistakes and provides targeted feedback; (3) after fine-tuning on this dataset, the final supervisor model is provided for real therapist training. The detailed experimental results of automated, human and downstream assessments demonstrate that models fine-tuned on our dataset MATE, can provide high-quality feedback according to the clinical guideline, showing significant potential for the therapist training scenario.

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.

Prototypical Progressive Alignment and Reweighting for Generalizable Semantic Segmentation

Generalizable semantic segmentation aims to perform well on unseen target domains, a critical challenge due to real-world applications requiring high generalizability. Class-wise prototypes, representing class centroids, serve as domain-invariant cues that benefit generalization due to their stability and semantic consistency. However, this approach faces three challenges. First, existing methods often adopt coarse prototypical alignment strategies, which may hinder performance. Second, naive prototypes computed by averaging source batch features are prone to overfitting and may be negatively affected by unrelated source data. Third, most methods treat all source samples equally, ignoring the fact that different features have varying adaptation difficulties. To address these limitations, we propose a novel framework for generalizable semantic segmentation: Prototypical Progressive Alignment and Reweighting (PPAR), leveraging the strong generalization ability of the CLIP model. Specifically, we define two prototypes: the Original Text Prototype (OTP) and Visual Text Prototype (VTP), generated via CLIP to serve as a solid base for alignment. We then introduce a progressive alignment strategy that aligns features in an easy-to-difficult manner, reducing domain gaps gradually. Furthermore, we propose a prototypical reweighting mechanism that estimates the reliability of source data and adjusts its contribution, mitigating the effect of irrelevant or harmful features (i.e., reducing negative transfer). We also provide a theoretical analysis showing the alignment between our method and domain generalization theory. Extensive experiments across multiple benchmarks demonstrate that PPAR achieves state-of-the-art performance, validating its effectiveness.

Precise Zero-Shot Pointwise Ranking with LLMs through Post-Aggregated Global Context Information

Recent advancements have successfully harnessed the power of Large Language Models (LLMs) for zero-shot document ranking, exploring a variety of prompting strategies. Comparative approaches like pairwise and listwise achieve high effectiveness but are computationally intensive and thus less practical for larger-scale applications. Scoring-based pointwise approaches exhibit superior efficiency by independently and simultaneously generating the relevance scores for each candidate document. However, this independence ignores critical comparative insights between documents, resulting in inconsistent scoring and suboptimal performance. In this paper, we aim to improve the effectiveness of pointwise methods while preserving their efficiency through two key innovations: (1) We propose a novel Global-Consistent Comparative Pointwise Ranking (GCCP) strategy that incorporates global reference comparisons between each candidate and an anchor document to generate contrastive relevance scores. We strategically design the anchor document as a query-focused summary of pseudo-relevant candidates, which serves as an effective reference point by capturing the global context for document comparison. (2) These contrastive relevance scores can be efficiently Post-Aggregated with existing pointwise methods, seamlessly integrating essential Global Context information in a training-free manner (PAGC). Extensive experiments on the TREC DL and BEIR benchmark demonstrate that our approach significantly outperforms previous pointwise methods while maintaining comparable efficiency. Our method also achieves competitive performance against comparative methods that require substantially more computational resources. More analyses further validate the efficacy of our anchor construction strategy.

Collapsing Sequence-Level Data-Policy Coverage via Poisoning Attack in Offline Reinforcement Learning

Offline reinforcement learning (RL) heavily relies on the coverage of pre-collected data over the target policy's distribution. Existing studies aim to improve data-policy coverage to mitigate distributional shifts, but overlook security risks from insufficient coverage, and the single-step analysis is not consistent with the multi-step decision-making nature of offline RL. To address this, we introduce the sequence-level concentrability coefficient to quantify coverage, and reveal its exponential amplification on the upper bound of estimation errors through theoretical analysis. Building on this, we propose the Collapsing Sequence-Level Data-Policy Coverage (CSDPC) poisoning attack. Considering the continuous nature of offline RL data, we convert state-action pairs into decision units, and extract representative decision patterns that capture multi-step behavior. We identify rare patterns likely to cause insufficient coverage, and poison them to reduce coverage and exacerbate distributional shifts. Experiments show that poisoning just 1% of the dataset can degrade agent performance by 90%. This finding provides new perspectives for analyzing and safeguarding the security of offline RL.

Step-Audio-AQAA: a Fully End-to-End Expressive Large Audio Language Model

Large Audio-Language Models (LALMs) have significantly advanced intelligent human-computer interaction, yet their reliance on text-based outputs limits their ability to generate natural speech responses directly, hindering seamless audio interactions. To address this, we introduce Step-Audio-AQAA, a fully end-to-end LALM designed for Audio Query-Audio Answer (AQAA) tasks. The model integrates a dual-codebook audio tokenizer for linguistic and semantic feature extraction, a 130-billion-parameter backbone LLM and a neural vocoder for high-fidelity speech synthesis. Our post-training approach employs interleaved token-output of text and audio to enhance semantic coherence and combines Direct Preference Optimization (DPO) with model merge to improve performance. Evaluations on the StepEval-Audio-360 benchmark demonstrate that Step-Audio-AQAA excels especially in speech control, outperforming the state-of-art LALMs in key areas. This work contributes a promising solution for end-to-end LALMs and highlights the critical role of token-based vocoder in enhancing overall performance for AQAA tasks.

A Survey of Automatic Evaluation Methods on Text, Visual and Speech Generations

Recent advances in deep learning have significantly enhanced generative AI capabilities across text, images, and audio. However, automatically evaluating the quality of these generated outputs presents ongoing challenges. Although numerous automatic evaluation methods exist, current research lacks a systematic framework that comprehensively organizes these methods across text, visual, and audio modalities. To address this issue, we present a comprehensive review and a unified taxonomy of automatic evaluation methods for generated content across all three modalities; We identify five fundamental paradigms that characterize existing evaluation approaches across these domains. Our analysis begins by examining evaluation methods for text generation, where techniques are most mature. We then extend this framework to image and audio generation, demonstrating its broad applicability. Finally, we discuss promising directions for future research in cross-modal evaluation methodologies.

T2I-Eval-R1: Reinforcement Learning-Driven Reasoning for Interpretable Text-to-Image Evaluation

The rapid progress in diffusion-based text-to-image (T2I) generation has created an urgent need for interpretable automatic evaluation methods that can assess the quality of generated images, therefore reducing the human annotation burden. To reduce the prohibitive cost of relying on commercial models for large-scale evaluation, and to improve the reasoning capabilities of open-source models, recent research has explored supervised fine-tuning (SFT) of multimodal large language models (MLLMs) as dedicated T2I evaluators. However, SFT approaches typically rely on high-quality critique datasets, which are either generated by proprietary LLMs-with potential issues of bias and inconsistency-or annotated by humans at high cost, limiting their scalability and generalization. To address these limitations, we propose T2I-Eval-R1, a novel reinforcement learning framework that trains open-source MLLMs using only coarse-grained quality scores, thereby avoiding the need for annotating high-quality interpretable evaluation rationale. Our approach integrates Group Relative Policy Optimization (GRPO) into the instruction-tuning process, enabling models to generate both scalar scores and interpretable reasoning chains with only easy accessible annotated judgment scores or preferences. Furthermore, we introduce a continuous reward formulation that encourages score diversity and provides stable optimization signals, leading to more robust and discriminative evaluation behavior. Experimental results on three established T2I meta-evaluation benchmarks demonstrate that T2I-Eval-R1 achieves significantly higher alignment with human assessments and offers more accurate interpretable score rationales compared to strong baseline methods.