Alibaba Group (China)

CRESiST replaces prefix-incremental generation in LLM-based simultaneous speech translation with an interleaved sequence of speech chunks and target tokens, so the KV cache is reused across the whole stream and no voice activity detection is needed to cut sentences. An RNN-T Actor samples read-or-write trajectories, a Qwen2.5-omni-7B Critic with Router and Refiner scores them on semantic consistency and latency, and reverse distillation feeds the decisions back to the Actor. On MuST-C and ACL 60/60, unsegmented streaming lands only 1–2 COMET points below the oracle-segmented setting.

CRESiST replaces prefix-incremental generation in LLM-based simultaneous speech translation with an interleaved sequence of speech chunks and target tokens, so the KV cache is reused across the whole stream and no voice activity detection is needed to cut sentences. An RNN-T Actor samples read-or-write trajectories, a Qwen2.5-omni-7B Critic with Router and Refiner scores them on semantic consistency and latency, and reverse distillation feeds the decisions back to the Actor. On MuST-C and ACL 60/60, unsegmented streaming lands only 1–2 COMET points below the oracle-segmented setting.

Visual verifiers have long returned only a true-or-false verdict, which cannot separate partially correct outputs from worse ones. OmniVerifier-C1 decomposes a complex prompt or environmental rule into fine-grained sub-questions, judges each one binarily, and turns the satisfied fraction into a continuous score, while a decoupled reinforcement-learning scheme optimizes sub-question judgments and global alignment scoring together. On ViVerBench, the Qwen3-VL-8B version trained with graded supervised fine-tuning plus graded reinforcement learning reaches an overall score of 0.684, above 0.671 for binary reinforcement learning alone.

Visual verifiers have long returned only a true-or-false verdict, which cannot separate partially correct outputs from worse ones. OmniVerifier-C1 decomposes a complex prompt or environmental rule into fine-grained sub-questions, judges each one binarily, and turns the satisfied fraction into a continuous score, while a decoupled reinforcement-learning scheme optimizes sub-question judgments and global alignment scoring together. On ViVerBench, the Qwen3-VL-8B version trained with graded supervised fine-tuning plus graded reinforcement learning reaches an overall score of 0.684, above 0.671 for binary reinforcement learning alone.

Real-time video-to-audio has been held back by two bottlenecks: visual frontends shrink their receptive fields to cut look-ahead, and causal diffusion backends denoise each patch in isolation, so both alignment and audio quality trail offline models. LiveFoley uses a trailing sliding window to turn off-the-shelf non-causal segment-level encoders into streaming modules without retraining, cutting frontend look-ahead to 0.08 s, and widens the backend denoising window toward history. On the VGGSound test set, LiveFoley-48k reaches a DeSync of 0.48, on par with offline MMAudio-L at 0.47, with 156 ms end-to-end latency on a single A800 GPU.

Real-time video-to-audio has been held back by two bottlenecks: visual frontends shrink their receptive fields to cut look-ahead, and causal diffusion backends denoise each patch in isolation, so both alignment and audio quality trail offline models. LiveFoley uses a trailing sliding window to turn off-the-shelf non-causal segment-level encoders into streaming modules without retraining, cutting frontend look-ahead to 0.08 s, and widens the backend denoising window toward history. On the VGGSound test set, LiveFoley-48k reaches a DeSync of 0.48, on par with offline MMAudio-L at 0.47, with 156 ms end-to-end latency on a single A800 GPU.

Liver CT diagnosis models usually predict disease labels directly from images, so they only recognize diseases seen in training. DbC splits diagnosis into two steps: it learns a set of radiological findings as dense maps over the scan, then defines each disease by the findings its guidelines require and composes the diagnosis map with a fixed, parameter-free soft-minimum operator. On multi-phase liver CT, for six diseases never seen in training, DbC localized lesions significantly better than every comparator, including models that align images with disease names, and kept its advantage on rare diseases from another institution and on a public external cohort.

Liver CT diagnosis models usually predict disease labels directly from images, so they only recognize diseases seen in training. DbC splits diagnosis into two steps: it learns a set of radiological findings as dense maps over the scan, then defines each disease by the findings its guidelines require and composes the diagnosis map with a fixed, parameter-free soft-minimum operator. On multi-phase liver CT, for six diseases never seen in training, DbC localized lesions significantly better than every comparator, including models that align images with disease names, and kept its advantage on rare diseases from another institution and on a public external cohort.

Knowledge improvement for large language models is moving from predefined incorporation pipelines to evaluation-guided updating, yet locating a weak spot does not say which intervention fits. Recursive Knowledge Improvement (RKI) organizes knowledge into graph-structured units and, each round, localizes weaknesses, estimates the model's operational knowledge state from diagnostic evidence and intervention history, selects an intervention, and re-evaluates the updated model. On HotpotEval and SafeInternalEval, RKI outperforms existing knowledge incorporation methods by 18.41 points on average, and analyses indicate knowledge states add guidance beyond weakness localization and can change across updates while the observed weakness persists.

Knowledge improvement for large language models is moving from predefined incorporation pipelines to evaluation-guided updating, yet locating a weak spot does not say which intervention fits. Recursive Knowledge Improvement (RKI) organizes knowledge into graph-structured units and, each round, localizes weaknesses, estimates the model's operational knowledge state from diagnostic evidence and intervention history, selects an intervention, and re-evaluates the updated model. On HotpotEval and SafeInternalEval, RKI outperforms existing knowledge incorporation methods by 18.41 points on average, and analyses indicate knowledge states add guidance beyond weakness localization and can change across updates while the observed weakness persists.

Agents use experimental feedback to revise data-engineering strategies, and across four tasks—math and code data selection, math reward design, and math curriculum learning—the selected strategies gain up to 13.34% absolute over generic baselines on AIME25 Pass@32 under fixed training conditions.

Agents use experimental feedback to revise data-engineering strategies, and across four tasks—math and code data selection, math reward design, and math curriculum learning—the selected strategies gain up to 13.34% absolute over generic baselines on AIME25 Pass@32 under fixed training conditions.

Multi-turn agent distillation now grounds teacher supervision in the interaction state, keeping only executed actions and observations, and the Qwen3-4B student surpasses its 32B teacher on ALFWorld Seen at 43.39%.

Multi-turn agent distillation now grounds teacher supervision in the interaction state, keeping only executed actions and observations, and the Qwen3-4B student surpasses its 32B teacher on ALFWorld Seen at 43.39%.