Reducing Rollout Bottlenecks in Reinforcement Learning with NVIDIA Dynamo
Reinforcement learning (RL) uses rewards to train models to solve problems and complete tasks more successfully. Inference drives this process. Workers generate...
Search headlines, summaries and articles to connect the dots.
Reinforcement learning (RL) uses rewards to train models to solve problems and complete tasks more successfully. Inference drives this process. Workers generate...
Most sellers do not need another AI product to learn. They need the tools and information they already use brought together, so they can see what is happening in their business and act on it.
Article URL: https://github.com/dynobox/dynobox Comments URL: https://news.ycombinator.com/item?id=50565029 Points: 1 # Comments: 0
Article URL: https://github.com/blackarck/intaiassist Comments URL: https://news.ycombinator.com/item?id=50572899 Points: 1 # Comments: 0
Article URL: https://kelu124.github.io/minus/ Comments URL: https://news.ycombinator.com/item?id=50563362 Points: 1 # Comments: 0
Article URL: https://github.com/PHONGUIT22/nostrpulse-full Comments URL: https://news.ycombinator.com/item?id=50572502 Points: 2 # Comments: 0
Article URL: https://publicrows.com Comments URL: https://news.ycombinator.com/item?id=50564744 Points: 1 # Comments: 0
Article URL: https://github.com/pug-sh/pug Comments URL: https://news.ycombinator.com/item?id=50576620 Points: 1 # Comments: 0
arXiv:2611.25126v1 Announce Type: new Abstract: Cross-modal knowledge distillation can incur negative transfer when source-modality supervision conflicts with the target modality's discriminative characteristics. We propose TALD, a two-stage framework for Transferability-Aware Logit Disentanglement in class-aligned logit space. Guided by task-dependent self-benefiting and cross-modal exclusivity criteria, a Residual Disentangler decomposes logits from a teacher and an independently trained target-modality reference student into modality-general and modality-specific components. A Sample-Adaptive Modulator regulates decomposition constraints and sample-specific fusion weights. The learned modality-general components are then adaptively fused to supervise the final student. Across five benchmarks spanning vision, audio, and text, TALD achieves the largest mean accuracy gain and highest positive-transfer frequency among the compared methods. Across ten transfer directions with five runs each, TALD improves over training without distillation in 84% of evaluations, compared with at most 70% for the baselines, while offering a favorable accuracy-efficiency trade-off. We plan to release the code.
arXiv:2611.26526v1 Announce Type: new Abstract: A benchmarking campaign provides the measurements behind an HPC performance study: code variants are built, swept over core or node counts, and measured repeatedly. Moving such a campaign to another supercomputer or rerunning it months later requires work that is not part of the study itself. Job scripts are adapted to each supercomputer, and what exactly ran is often reconstructed afterwards. We present Thinflow, a small Bash engine for organizing HPC benchmarking campaigns as rerunnable directory trees of tasks and runs. A task is a directory with a short shell script next to the code. For each sweep, one command builds the code, runs the benchmarks, and plots the results. Thinflow derives the execution order from declared dependencies and writes it to a plain-text manifest. Per-supercomputer profiles submit the runs to a workload manager such as Slurm. Each run writes a provenance record of its environment, hardware, and parameters. The engine ships inside the campaign directory, so a campaign stays rerunnable without installing a separate tool and, once completed, may serve as a paper artifact. In a case study, we ran one campaign with identical task files on three Slurm partitions of two supercomputers, on up to 16 nodes. This strong-scaling study of matrix multiplication in C/OpenMP, C/MPI, and Julia/MPI comprises 37 task files with 217 non-comment lines. Its 455 benchmark runs recorded what differed per partition, chiefly profiles and command-line parameters. This suggests that such a campaign needs little structure beyond plain shell files.
Article URL: https://chevyclutchfoot.substack.com/p/thought-process-before-and-after Comments URL: https://news.ycombinator.com/item?id=50572638 Points: 24 # Comments: 11
Kimi Code CLI v2.7.0 adds a native Amazon Bedrock provider, so you can connect the open source coding agent to models like Kimi K3.1 Pro using your standard AWS credential chain (named profiles or SSO) instead of a separate API key. This post shows how to configure the provider and send your first request.
arXiv:2611.25000v1 Announce Type: new Abstract: Model merging combines separately fine-tuned models by adding their task vectors to a shared backbone. Interference is typically assessed only after constructing and evaluating a candidate merge, making coefficient search expensive. We show that a first-order directional derivative of task A's outputs along task B's vector can predict interference, but only when evaluated at the appropriate model. At the receiver model, this derivative forecasts nearby merges per task and per example, halves the accuracy error of a no-change baseline, and identifies most examples whose predictions flip. For full multi-task merges, a single evaluation is insufficient; two evaluations along the merge path are needed. At the shared pretrained backbone, however, the derivative is nearly orthogonal to realized changes in our vision experiments and only partly predictive in LoRA language-model pilots. Controls link this failure to displacement along the receiver's own task vector, not simply parameter distance; the mismatch appears within the first few dozen fine-tuning steps. Freezing receiver-local responses produces a surrogate affine in the merge coefficients. We use it to construct LR-Merge, a label-free correction that optimizes coefficients through convex teacher matching on unlabeled probes, without backpropagation through merged networks. LR-Merge improves all tested closed-form, isotropic, and sparsified base mergers, with gains generally increasing as more tasks are merged. Using pretrained rather than receiver-local responses instead degrades the base merge. LR-Merge matches exact teacher matching through a merged network, which takes roughly three times longer to converge and performs similarly at an equal compute budget. Once built, the surrogate answers new coefficient and task-subset queries in seconds.
clickhousectl v0.6.0 finds projects the way Git does, connects to Cloud without a password, adds search, shell completions and feedback, and verifies every install.
A strong AI agent governance strategy
arXiv:2611.23883v1 Announce Type: cross Abstract: Ising machines, dedicated accelerators built around the dynamics of Ising-like spin models, have emerged as a leading platform for combinatorial optimization and probabilistic sampling. The problems that motivate them require millions of spins, far beyond the on-chip capacity of any single device, so a practical Ising machine must distribute its computation across many devices. Prior work reports throughput at a single operating point, leaving the binding resource unidentified and platform crossovers measurable only by experiment; no calibrated cross-platform cost model of distributed Ising machines has been reported, nor have systematic techniques to accelerate their Markov chain Monte Carlo (MCMC) workloads. This work presents a calibrated cost model that decomposes digital Ising sweep time into compute, memory, synchronization, and inter-device communication terms, validated with 4.8% mean error across three GPU generations, FPGA and ASIC up to 1 million spins. Guided by the model, we address each term in turn and propose: (i) a data placement strategy, worth up to 4.2$\times$ speedup; (ii) multi-rate and bounded-staleness schedules that relax the barrier that Gibbs sampling imposes, cutting time-to-solution (TTS) by up to 3.16$\times$ on three benchmarks; and (iii) a phase-offset streaming dataflow that hides boundary exchange behind interior computation whenever a link satisfies a latency-transparency criterion, yielding up to 1.25$\times$/1.74$\times$/2.01$\times$ on an H200 pair, a ZCU106 pair, and a four-board XEM8320 $2{\times}2$ mesh. The streaming dataflow is also implemented on a fabricated four-chip 28 nm ASIC sustaining 28.1 Gflips/s at 4.81 pJ/flip.
Coding agents can now use the Vercel CLI's help search feature to discover commands by describing tasks in plain language, with support for API endpoints, JSON output, and error tolerance. The feature includes a built-in skill prompting agents to start with search, and no search queries are logged.
Review logs, annotate data, and build datasets with LangSmith's Data Annotation Queues. Improve LLM app performance from prototype to production.
Article URL: https://www.lawyerscommittee.org/lawyers-committee-files-lawsuit-on-behalf-of-black-students-detained-and-searched-due-to-ai-alert-falsely-identifying-a-bag-of-chips-as-a-firearm/ Comments URL: https://news.ycombinator.com/item?id=50554641 Points: 27 # Comments: 0
Many enterprises run a centralized network model: one account owns every VPC and a central firewall inspects all traffic. This post shows a tested pattern for connecting Amazon Bedrock AgentCore Gateway to private APIs and MCP servers in that model, with the VPC Lattice resource gateway in the network account so every hop crosses the firewall.