Searcharxiv⌕ Search

arXiv subjects

Jiayang Chen

Publications and source records attributed to Jiayang Chen.

16 recordsLinked to original sources

Scaling RNA Foundation Models through Iterative Refinement

Biological function in RNA is encoded across scales, from local sequence motifs to long-range structural dependencies. Conventional RNA foundation models couple parameter capacity and inference depth within a fixed architecture, prescribing a fixed computation graph for each inference. Here we present Eco-RNA, an 80M-parameter RNA foundation model that makes inference depth adjustable through learned iterative refinement. Its shared, input-conditioned Transformer core refines sequence representations through repeated application. The iteration count is selected after training and controls effective depth at fixed model parameters. We train across recurrent depths with masked reconstruction, state regularization, and a one-sided loss that anchors deeper reconstruction to a shallow reference. On the 70-assay RNAGym fitness benchmark, increasing refinement from one to four iterations raises aggregate absolute Spearman correlation from 0.242 to 0.286, with the gain retained through eight iterations. At four iterations, Eco-RNA surpasses the current SOTA Evo 2 40B on all three aggregate fitness metrics despite 500x fewer parameters. Its iteration configurations advance the compute-performance Pareto frontier among the evaluated models. It also outperforms the evaluated baselines in non-coding RNA family classification, RNA secondary-structure prediction, and chemical probing profile prediction. These results establish internal inference computation as an effective scaling dimension for RNA foundation models.

cs.LG↗

Self-Evolving Dual-Role Deliberation for Medication Recommendation from Clinical Records

Recommending medications from electronic health records (EHRs) is a safety-critical task that demands therapeutic coverage, explicit safety constraints, traceable justification, and continual adaptivity. Supervised neural models freeze at deployment and cannot reason about new phenotypes, while existing LLM pipelines are single-pass and lack a cross-case improvement loop. We propose SEMPA, a training-free agentic framework that couples a Proposer with a Critic in multi-round adversarial deliberation, grounds candidates with molecular-aware retrieval, contrastively evolves both agents' natural-language policies under a semantic-anchor constraint, and enforces rule-based DDI and contraindication checks. On 100 MIMIC-IV encounters, SEMPA reaches 0.3116 Jaccard, 0.7295 F1, and a 0.0172 DDI rate, improving F1 by 57.7% and cutting DDI by 77.5% over the strongest supervised baseline. Two LLM judges rank SEMPA highest, and six-backbone studies confirm that gains derive from the framework rather than a specific model.

cs.CL↗

CONDUIT: A Unified Residual-Stream Restoration Framework for KV Cache Reuse in Vision-Language Models

Vision-language models (VLMs) often answer new questions about recurring visual content, where reusing the key-value (KV) cache can avoid re-encoding expensive visual prefixes. Exact-prefix reuse, however, fails when the same visual content appears under a changed prefix. Selective recomputation can recover quality under a small visual-token budget, but only when the right stale tokens are refreshed. Raw-attention selection can waste budget on high-attention tokens with small value-norm proxy scores and on query-irrelevant images. To address these failure modes, we propose CONDUIT, a training-free refresh policy that unifies single- and multi-image reuse as residual-stream restoration. Building on norm-weighted attention, CONDUIT ranks cached visual tokens using cached-key query attention and an accessible pre-output cached-value-norm proxy, then applies empirical image-level relevance amplification before one global selection. With one image, the coefficient is one and the rule reduces to intra-image token selection. The method preserves model architecture and weights, adding only a single query-conditioned scoring pass at inference. At a 10% refresh budget, CONDUIT achieves 97.0-99.5% of the corresponding full-prefill five-dataset average across three VLM backbones and leads budgeted methods on average; on the MMLongBench-Doc latency subset, it uses 13.5% of full-prefill FLOPs and achieves a 2.99x time-to-first-token speedup.

cs.CL↗

PowerSlider: Exploiting Phase Asymmetry for LLM Serving under Demand Response

AI inference clusters are increasingly constrained by instantaneous power, not just energy: grid operators condition new capacity on demand response, imposing time-varying power caps. Existing LLM serving systems optimize a static energy objective or shed fixed priority tiers under load; either way, goodput collapses when the power envelope moves. An LLM pipeline is not a uniform load: compute-bound prefill loses throughput almost linearly with GPU frequency, memory-bound answer decode sustains it down to $0.57\times$ nominal, and reasoning's thinking phase couples KV-cache capacity to scheduling -- so a cap should be steered to where each watt costs the least performance. PowerSlider does so with a new Flex SLO contract that turns bounded user slack into an optimization constraint, prefill--think--answer disaggregation exposing per-stage frequency and KV control, and a Karush--Kuhn--Tucker (KKT) online solver re-solving within 7.7 ms of every cap change, backed by a consolidated fail-safe that power-gates drained instances when DVFS bottoms out on static power. On SGLang with production traces, \sys{} sustains 78.3\% online goodput at a 30\% cap reduction versus 47.6\% for the best of five baselines ($1.64\times$), holds latency-critical tails within $1.3\times$ of nominal (baselines: $2.3$--$6\times$, up to $12\times$), and delivers 92\% mean goodput through a replayed CAISO grid-emergency day bottoming at $0.41\times$ (54\% at the trough; every baseline below 7\%).

cs.DC↗

Beyond Prediction: Tail-Aware Scheduling for LLM Inference

LLM serving exhibits extreme length variability, making size-based scheduling difficult in practice. Recent LLM schedulers approximate SJF/SRPT using predicted decode lengths or ranks and primarily report mean-centric metrics such as TTFT and TBT. We show that these prediction-driven policies can be fragile under distribution shifts, bursty arrivals, and GPU memory pressure, while offering limited control over the tail latency (P90-P99) that dominates user experience, even with perfect decode-length knowledge. We introduce a distribution-aware, prediction-free scheduling framework that replaces explicit length prediction with soft priority boosting driven by lightweight statistical signals. Our design co-optimizes scheduling and cache-aware preemption to account for memory-coupled decode dynamics across workload mixes. Evaluated on production and open-source traces, our method reduces P99 TTLT by up to 35-50% relative to SRPT with perfect length knowledge and reduces TTFT by 34-47% across workloads, including reasoning-heavy and chat-heavy tasks. These results demonstrate a robust alternative for optimizing tail latency in online LLM serving.

cs.LG↗

DeepEN: A Deep Reinforcement Learning Framework for Personalized Enteral Nutrition in Critical Care

Objective: Enteral nutrition (EN) delivery in the ICU remains suboptimal due to limited personalization and uncertainty regarding appropriate calorie, protein, and fluid targets under dynamic metabolic demands. We introduce DeepEN, a reinforcement learning (RL) framework for personalized EN optimization using electronic health record data. Methods: DeepEN was trained on over 11,000 ICU patients from MIMIC-IV to generate 4-hourly, patient-specific caloric, protein, and fluid targets. The state representation incorporated demographics, comorbidities, vital signs, laboratory values, and recent interventions. A physiologically aligned reward framework balanced biomarker stability with long-term survival. Policy learning employed a dueling double deep Q-network with Conservative Q-Learning regularization to enable safe offline training. Results: DeepEN achieved the highest estimated policy value ($V^π= 9.48$) and the lowest calibrated mortality (18.8 +/- 1.0%), representing a 4.0 percentage-point absolute reduction compared with clinician practice (22.8%). The policy also demonstrated superior metabolic stability, achieving the highest proportion of glucose, phosphate, and sodium values within target range. Furthermore, deviation from the DeepEN policy was independently associated with increased mortality and biomarker instability, whereas deviation from a random policy showed no such association. Interpretability analyses further indicated that recommendations were conditioned on physiologically relevant markers of organ function and metabolic status rather than static dosing heuristics. Conclusion: DeepEN demonstrates the feasibility of conservative offline RL for safe, individualized EN optimization, highlighting the potential of data-driven personalization to complement guideline-based approaches in critical care.

cs.LG↗

Investigating Data Pruning for Pretraining Biological Foundation Models at Scale

Biological foundation models (BioFMs), pretrained on large-scale biological sequences, have recently shown strong potential in providing meaningful representations for diverse downstream bioinformatics tasks. However, such models often rely on millions to billions of training sequences and billions of parameters, resulting in prohibitive computational costs and significant barriers to reproducibility and accessibility, particularly for academic labs. To address these challenges, we investigate the feasibility of data pruning for BioFM pretraining and propose a post-hoc influence-guided data pruning framework tailored to biological domains. Our approach introduces a subset-based self-influence formulation that enables efficient estimation of sample importance at low computational cost, and builds upon it two simple yet effective selection strategies, namely Top-k Influence (Top I) and Coverage-Centric Influence (CCI). We empirically validate our method on two representative BioFMs, RNA-FM and ESM-C. For RNA, our framework consistently outperforms random selection baselines under an extreme pruning rate of over 99 percent, demonstrating its effectiveness. Furthermore, we show the generalizability of our framework on protein-related tasks using ESM-C. In particular, our coreset even outperforms random subsets that are ten times larger in both RNA and protein settings, revealing substantial redundancy in biological sequence datasets. These findings underscore the potential of influence-guided data pruning to substantially reduce the computational cost of BioFM pretraining, paving the way for more efficient, accessible, and sustainable biological AI research.

cs.LG↗

R$^2$: A LLM Based Novel-to-Screenplay Generation Framework with Causal Plot Graphs

Automatically adapting novels into screenplays is important for the TV, film, or opera industries to promote products with low costs. The strong performances of large language models (LLMs) in long-text generation call us to propose a LLM based framework Reader-Rewriter (R$^2$) for this task. However, there are two fundamental challenges here. First, the LLM hallucinations may cause inconsistent plot extraction and screenplay generation. Second, the causality-embedded plot lines should be effectively extracted for coherent rewriting. Therefore, two corresponding tactics are proposed: 1) A hallucination-aware refinement method (HAR) to iteratively discover and eliminate the affections of hallucinations; and 2) a causal plot-graph construction method (CPC) based on a greedy cycle-breaking algorithm to efficiently construct plot lines with event causalities. Recruiting those efficient techniques, R$^2$ utilizes two modules to mimic the human screenplay rewriting process: The Reader module adopts a sliding window and CPC to build the causal plot graphs, while the Rewriter module generates first the scene outlines based on the graphs and then the screenplays. HAR is integrated into both modules for accurate inferences of LLMs. Experimental results demonstrate the superiority of R$^2$, which substantially outperforms three existing approaches (51.3%, 22.6%, and 57.1% absolute increases) in pairwise comparison at the overall win rate for GPT-4o.

cs.AI↗

Accurate RNA 3D structure prediction using a language model-based deep learning approach

Accurate prediction of RNA three-dimensional (3D) structure remains an unsolved challenge. Determining RNA 3D structures is crucial for understanding their functions and informing RNA-targeting drug development and synthetic biology design. The structural flexibility of RNA, which leads to scarcity of experimentally determined data, complicates computational prediction efforts. Here, we present RhoFold+, an RNA language model-based deep learning method that accurately predicts 3D structures of single-chain RNAs from sequences. By integrating an RNA language model pre-trained on ~23.7 million RNA sequences and leveraging techniques to address data scarcity, RhoFold+ offers a fully automated end-to-end pipeline for RNA 3D structure prediction. Retrospective evaluations on RNA-Puzzles and CASP15 natural RNA targets demonstrate RhoFold+'s superiority over existing methods, including human expert groups. Its efficacy and generalizability are further validated through cross-family and cross-type assessments, as well as time-censored benchmarks. Additionally, RhoFold+ predicts RNA secondary structures and inter-helical angles, providing empirically verifiable features that broaden its applicability to RNA structure and function studies.

q-bio.QM↗

Detection of carbon monoxide using a polarization-multiplexed erbium dual-comb fiber laser

We present a compact, reliable, and robust free-running all-polarization-maintaining erbium (Er) single-cavity dual-comb laser generated via polarization multiplexing with gain sharing. Polarization multiplexing exploits the fast and slow axes of the fiber, while modelocking is achieved through a nonlinear amplifying loop mirror scheme using readily available components. The laser operates at a repetition rate of around 74.74 MHz with a tuning capability in the difference in repetition rates from 500 Hz to 200 kHz. This tunability makes the system more flexible for dual-comb spectroscopy experiments. Consequently, using this laser, we demonstrated a proof-of-principle dual-comb spectroscopy of carbon monoxide (CO), operating without any active stabilization.

physics.optics↗

An Invertible All-optical Logic Gate on Chip

We demonstrate an invertible all-optical gate on chip, with the roles of control and signal switchable by slightly adjusting their relative arrival time at the gate. It is based on quantum Zeno blockade driven by sum-frequency generation in a periodic-poled lithium niobate microring resonator. For two nearly-identical nanosecond pulses, the later arriving pulse is modulated by the earlier arriving one, resulting in 2.4 and 3.9 power extinction between the two, respectively, when their peak power is 1 mW and 2 mW. Our results, while to be improved and enriched, herald a new paradigm of logical gates and circuits for exotic applications.

physics.optics↗

Parametric All-Optical Modulation on Chip

We demonstrate parametric all-optical modulation in a periodically-poled lithium niobate microring resonator on chip. It employs quantum Zeno blockade between two distinct waves, a signal and a pump, through their sum-frequency generation at a large per-photon efficiency of 8.2 MHz. With nanosecond pump pulses at 6 mW peak power, 85.7% modulation extinction is observed, marking over 30~times efficiency improvement across various previous implementations. With only 2 mW pump peak power, 43.0% modulation extinction is observed for a doubly-stronger signal at 4 mW. This demonstrates, for the first time, that optical transistors with cascadability and fan-out are possible with just parametric nonlinear optics. These results, together with inherent advantages in such photonic integrated circuits, open the door to scalable technology for all-optical and quantum information processing.

physics.optics↗

Enhancing the Protein Tertiary Structure Prediction by Multiple Sequence Alignment Generation

The field of protein folding research has been greatly advanced by deep learning methods, with AlphaFold2 (AF2) demonstrating exceptional performance and atomic-level precision. As co-evolution is integral to protein structure prediction, AF2's accuracy is significantly influenced by the depth of multiple sequence alignment (MSA), which requires extensive exploration of a large protein database for similar sequences. However, not all protein sequences possess abundant homologous families, and consequently, AF2's performance can degrade on such queries, at times failing to produce meaningful results. To address this, we introduce a novel generative language model, MSA-Augmenter, which leverages protein-specific attention mechanisms and large-scale MSAs to generate useful, novel protein sequences not currently found in databases. These sequences supplement shallow MSAs, enhancing the accuracy of structural property predictions. Our experiments on CASP14 demonstrate that MSA-Augmenter can generate de novo sequences that retain co-evolutionary information from inferior MSAs, thereby improving protein structure prediction quality on top of strong AF2.

q-bio.QM↗

Interpretable RNA Foundation Model from Unannotated Data for Highly Accurate RNA Structure and Function Predictions

Non-coding RNA structure and function are essential to understanding various biological processes, such as cell signaling, gene expression, and post-transcriptional regulations. These are all among the core problems in the RNA field. With the rapid growth of sequencing technology, we have accumulated a massive amount of unannotated RNA sequences. On the other hand, expensive experimental observatory results in only limited numbers of annotated data and 3D structures. Hence, it is still challenging to design computational methods for predicting their structures and functions. The lack of annotated data and systematic study causes inferior performance. To resolve the issue, we propose a novel RNA foundation model (RNA-FM) to take advantage of all the 23 million non-coding RNA sequences through self-supervised learning. Within this approach, we discover that the pre-trained RNA-FM could infer sequential and evolutionary information of non-coding RNAs without using any labels. Furthermore, we demonstrate RNA-FM's effectiveness by applying it to the downstream secondary/3D structure prediction, SARS-CoV-2 genome structure and evolution prediction, protein-RNA binding preference modeling, and gene expression regulation modeling. The comprehensive experiments show that the proposed method improves the RNA structural and functional modelling results significantly and consistently. Despite only being trained with unlabelled data, RNA-FM can serve as the foundational model for the field.

q-bio.QM↗

Photon Conversion in Thin-film Lithium Niobate Nanowaveguides: A Noise Analysis

Wavelength transduction of single-photon signals is indispensable to networked quantum applications, particularly those incorporating quantum memories. Lithium niobate nanophotonic devices have demonstrated favorable linear, nonlinear, and electro-optical properties to deliver this crucial function while offering superiror efficiency, integrability, and scalability. Yet, their quantum noise level--an crucial metric for any single-photon based application--has yet to be understood. In this work, we report the first study with the focus on telecom to near-visible conversion driven by a telecom pump of small detuning, for practical considerations in distributed quantum processing over fiber networks. Our results find the noise level to be on the order of $10^{-4}$ photons per time-frequency mode for high conversion, allowing faithful pulsed operations. Through carefully analyzing the origins of such noise and each's dependence on the pump power and wavelength detuning, we have also identified a formula for noise suppression to $10^{-5}$ photons per mode. Our results assert a viable, low-cost, and modular approach to networked quantum processing and beyond using lithium niobate nanophotonics.

physics.optics↗

Anti-Parity-Time Symmetry in Passive Nanophotonics

Parity-time (PT) symmetry in non-Hermitian optical systems promises distinct optical effects and applications not found in conservative optics. Its counterpart, anti-PT symmetry, subscribes another class of intriguing optical phenomena and implies complementary techniques for exotic light manipulation. Despite exciting progress, so far anti-PT symmetry has only been realized in bulky systems or with optical gain. Here, we report an on-chip realization of non-Hermitian optics with anti-PT symmetry, by using a fully-passive, nanophotonic platform consisting of three evanescently coupled waveguides. By depositing a metal film on the center waveguide to introduce strong loss, an anti-PT system is realized. Using microheaters to tune the waveguides' refractive indices, striking behaviors are observed such as equal power splitting, synchronized amplitude modulation, phase-controlled dissipation, and transition from anti-PT symmetry to its broken phase. Our results highlight exotic anti-Hermitian nanophotonics to be consolidated with conventional circuits on the same chip, whereby valuable chip devices can be created for quantum optics studies and scalable information processing.

physics.optics↗