arXiv subjects

Jing Pan

Publications and source records attributed to Jing Pan.

At least 19 recordsLinked to original sources

arXiv2026

Manifold partitioning induced sequential optical reasoning and decision framework for photonic computing

Real-world data are intrinsically embedded in highly entangled manifolds, making the extraction of separable representations a central challenge for artificial intelligent (AI) systems. While optical neural networks (ONNs) offer ultrafast and energy-efficient data processing, their capacity is constrained by limited physical depth. Here, we introduce a sequential optical reasoning and decision (SORD) framework, an architecture that performs time-sequenced hierarchical inference by decomposing global tasks into coarse-to-fine steps via geometry-guided data partitioning. At each step, SORD executes small reasoning via dynamic operator selection, effectively reducing the overall task complexity without scaling up physical architecture. Experimentally, SORD enables a single-layer diffractive ONN to achieve otherwise intractable 100-class optical fiber speckle classification with 94% accuracy and a system energy efficiency of 23.3 TOPS/W. This high-fidelity recognition is further examined in a human-machine interface, featuring real-time interactive all-optical sensing. Overall, our work establishes a scalable and hardware-efficient approach to expanding the effective expressivity of compact photonic AI systems, and may advance their deployment in applications requiring real-time sensing, inference, and control.

physics.optics↗

arXiv2026

Equipping LLM with Directional Multi-Talker Speech Understanding Capabilities

Recent studies have demonstrated that prompting large language models (LLM) with audio encodings enables effective speech understanding capabilities. However, most speech LLMs are trained on single-channel, single-talker data, which makes it challenging to directly apply them to multi-talker and multi-channel speech understanding task. In this work, we present a comprehensive investigation on how to enable directional multi-talker speech understanding capabilities for LLMs, specifically in smart glasses usecase. We propose two novel approaches to integrate directivity into LLMs: (1) a cascaded system that leverages a source separation front-end module, and (2) an end-to-end system that utilizes serialized output training. All of the approaches utilize a multi-microphone array embedded in smart glasses to optimize directivity interpretation and processing in a streaming manner. Experimental results demonstrate the efficacy of our proposed methods in endowing LLMs with directional speech understanding capabilities, achieving strong performance in both speech recognition and speech translation tasks.

cs.CL↗

arXiv2026

WADBERT: Dual-channel Web Attack Detection Based on BERT Models

Web attack detection is the first line of defense for securing web applications, designed to preemptively identify malicious activities. Deep learning-based approaches are increasingly popular for their advantages: automatically learning complex patterns and extracting semantic features from HTTP requests to achieve superior detection performance. However, existing methods are less effective in embedding irregular HTTP requests, even failing to model unordered parameters and achieve attack traceability. In this paper, we propose an effective web attack detection model, named WADBERT. It achieves high detection accuracy while enabling the precise identification of malicious parameters. To this end, we first employ Hybrid Granularity Embedding (HGE) to generate fine-grained embeddings for URL and payload parameters. Then, URLBERT and SecBERT are respectively utilized to extract their semantic features. Further, parameter-level features (extracted by SecBERT) are fused through a multi-head attention mechanism, resulting in a comprehensive payload feature. Finally, by feeding the concatenated URL and payload features into a linear classifier, a final detection result is obtained. The experimental results on CSIC2010 and SR-BH2020 datasets validate the efficacy of WADBERT, which respectively achieves F1-scores of 99.63% and 99.50%, and significantly outperforms state-of-the-art methods.

cs.CR↗

arXiv2025

A Large Scale Heterogeneous Treatment Effect Estimation Framework and Its Applications of Users' Journey at Snap

Heterogeneous Treatment Effect (HTE) and Conditional Average Treatment Effect (CATE) models relax the assumption that treatment effects are the same for every user. We present a large scale industrial framework for estimating HTE using experimental data from hundreds of millions of Snapchat users. By combining results across many experiments, the framework uncovers latent user characteristics that were previously unmeasurable and produces stable treatment effect estimates at scale. We describe the core components that enabled this system, including experiment selection, base learner design, and incremental training. We also highlight two applications: user influenceability to ads and user sensitivity to ads. An online A/B test using influenceability scores for targeting showed an improvement on key business metrics that is more than six times larger than what is typically considered significant.

cs.LG↗

arXiv2025

A Mathematical Model of Hematopoiesis during Systemic Infection

Chronic critical illness (CCI) is a disease state in which, following an initial insult, a patient neither recovers nor dies but instead remains in a state of critical illness. CCI is characterized by prolonged organ dysfunction, weight loss, and persistent increased vulnerability to infection. Recent data has shown that patients with CCI generally exhibit persistent immune dysfunction, characterized by prolonged elevation of specific pro-inflammatory cytokines. In this paper, we introduce a host response model that couples hematopoiesis dynamics with immune response to infection. Specifically, we incorporate the reactions between pro-inflammatory and anti-inflammatory signals with specific hematopoietic stem cell compartments with a reduced model of acute inflammation. We found that a maladaptive hematopoietic response to pathogenic insult is able to qualitatively reproduce similar behavior to that seen in CCI patients, namely the presence of a persistent, elevated level of pro-inflammatory cytokines. This suggests that maladaptive hematopoietic responses in vivo may play a role in the development of CCI.

q-bio.CB↗

arXiv2025

Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs

We introduce Phi-4-Mini and Phi-4-Multimodal, compact yet highly capable language and multimodal models. Phi-4-Mini is a 3.8-billion-parameter language model trained on high-quality web and synthetic data, significantly outperforming recent open-source models of similar size and matching the performance of models twice its size on math and coding tasks requiring complex reasoning. This achievement is driven by a carefully curated synthetic data recipe emphasizing high-quality math and coding datasets. Compared to its predecessor, Phi-3.5-Mini, Phi-4-Mini features an expanded vocabulary size of 200K tokens to better support multilingual applications, as well as group query attention for more efficient long-sequence generation. Phi-4-Multimodal is a multimodal model that integrates text, vision, and speech/audio input modalities into a single model. Its novel modality extension approach leverages LoRA adapters and modality-specific routers to allow multiple inference modes combining various modalities without interference. For example, it now ranks first in the OpenASR leaderboard to date, although the LoRA component of the speech/audio modality has just 460 million parameters. Phi-4-Multimodal supports scenarios involving (vision + language), (vision + speech), and (speech/audio) inputs, outperforming larger vision-language and speech-language models on a wide range of tasks. Additionally, we experiment to further train Phi-4-Mini to enhance its reasoning capabilities. Despite its compact 3.8-billion-parameter size, this experimental version achieves reasoning performance on par with or surpassing significantly larger models, including DeepSeek-R1-Distill-Qwen-7B and DeepSeek-R1-Distill-Llama-8B.

cs.CL↗

arXiv2024

Target word activity detector: An approach to obtain ASR word boundaries without lexicon

Obtaining word timestamp information from end-to-end (E2E) ASR models remains challenging due to the lack of explicit time alignment during training. This issue is further complicated in multilingual models. Existing methods, either rely on lexicons or introduce additional tokens, leading to scalability issues and increased computational costs. In this work, we propose a new approach to estimate word boundaries without relying on lexicons. Our method leverages word embeddings from sub-word token units and a pretrained ASR model, requiring only word alignment information during training. Our proposed method can scale-up to any number of languages without incurring any additional cost. We validate our approach using a multilingual ASR model trained on five languages and demonstrate its effectiveness against a strong baseline.

cs.CL↗

arXiv2024

CogniDual Framework: Self-Training Large Language Models within a Dual-System Theoretical Framework for Improving Cognitive Tasks

Cognitive psychology investigates perception, attention, memory, language, problem-solving, decision-making, and reasoning. Kahneman's dual-system theory elucidates the human decision-making process, distinguishing between the rapid, intuitive System 1 and the deliberative, rational System 2. Recent advancements have positioned large language Models (LLMs) as formidable tools nearing human-level proficiency in various cognitive tasks. Nonetheless, the presence of a dual-system framework analogous to human cognition in LLMs remains unexplored. This study introduces the \textbf{CogniDual Framework for LLMs} (CFLLMs), designed to assess whether LLMs can, through self-training, evolve from deliberate deduction to intuitive responses, thereby emulating the human process of acquiring and mastering new information. Our findings reveal the cognitive mechanisms behind LLMs' response generation, enhancing our understanding of their capabilities in cognitive psychology. Practically, self-trained models can provide faster responses to certain queries, reducing computational demands during inference.

cs.CL↗

arXiv2024

BTBR: A Bayesian-Theory-Driven Probabilistic-Fuzzy Framework for Implicit Bias Removal in Large Language Models

Large language models (LLMs) may encode biased associations from heterogeneous training corpora that are not immediately visible under ordinary prompting, but can surface when the model is steered toward particular demographic personas. Such behavior often manifests not as explicit toxic output, but as systematic performance differences across semantically equivalent tasks, making the resulting bias difficult to detect and mitigate. To address this issue, we formalize the implicit bias problem as persona-induced performance disparity and argue that bias evidence should be treated as a graded signal rather than a binary label. Motivated by this observation, we model biased knowledge as a fuzzy subset equipped with an explicit membership function that reflects the strength of bias evidence for each candidate example. Building on this formulation, we propose Bayesian-Theory-based Bias Removal (BTBR), a hybrid probabilistic-fuzzy framework for identifying and removing latent bias traces from model parameters. BTBR first performs likelihood-ratio screening to measure how strongly candidate samples align with a target biased persona, then converts high-membership samples into structured knowledge triples, and finally applies targeted model editing with a lightweight fuzzy rule scheduler to reduce collateral performance degradation under high entanglement risk. Extensive experiments across multiple bias sources, tasks, model families and editing backends show that BTBR consistently reduces persona-induced performance gaps while preserving general reasoning ability. These results demonstrate that combining probabilistic evidence with fuzzy degree modeling provides an effective and practical approach for mitigating implicit bias in large language models.

cs.CL↗

arXiv2024

Photonic Landau levels in a high-dimensional frequency-degenerate cavity

Topological orders emerge in both microscopic quantum dynamics and macroscopic materials as a fundamental principle to characterize intricate properties in nature with vital significance, for instance, the Landau levels of electron systems in magnetic field. Whilst, recent advances of synthetic photonic systems enable generalized concepts of Landau levels across fermionic and bosonic systems, extending the modern physical frontier. However, the controls of Landau levels of photons were only confined in complex artificial metamaterials or multifolded cavities. Here, we exploit advanced structured light laser technology and propose the theory of high-dimensional frequency-degeneracy, which enables photonic Landau level control in a linear open laser cavity with simple displacement tuning of intracavity elements. This work not only create novel structured light with new topological effects but also provides broad prospects for Bose-analogue quantum Hall effects and topological physics.

physics.optics↗

arXiv2024

WavLLM: Towards Robust and Adaptive Speech Large Language Model

The recent advancements in large language models (LLMs) have revolutionized the field of natural language processing, progressively broadening their scope to multimodal perception and generation. However, effectively integrating listening capabilities into LLMs poses significant challenges, particularly with respect to generalizing across varied contexts and executing complex auditory tasks. In this work, we introduce WavLLM, a robust and adaptive speech large language model with dual encoders, and a prompt-aware LoRA weight adapter, optimized by a two-stage curriculum learning approach. Leveraging dual encoders, we decouple different types of speech information, utilizing a Whisper encoder to process the semantic content of speech, and a WavLM encoder to capture the unique characteristics of the speaker's identity. Within the curriculum learning framework, WavLLM first builds its foundational capabilities by optimizing on mixed elementary single tasks, followed by advanced multi-task training on more complex tasks such as combinations of the elementary tasks. To enhance the flexibility and adherence to different tasks and instructions, a prompt-aware LoRA weight adapter is introduced in the second advanced multi-task training stage. We validate the proposed model on universal speech benchmarks including tasks such as ASR, ST, SV, ER, and also apply it to specialized datasets like Gaokao English listening comprehension set for SQA, and speech Chain-of-Thought (CoT) evaluation set. Experiments demonstrate that the proposed model achieves state-of-the-art performance across a range of speech tasks on the same model size, exhibiting robust generalization capabilities in executing complex tasks using CoT approach. Furthermore, our model successfully completes Gaokao tasks without specialized training. The codes, models, audio, and Gaokao evaluation set can be accessed at \url{aka.ms/wavllm}.

cs.CL↗

arXiv2023

COSMIC: Data Efficient Instruction-tuning For Speech In-Context Learning

We present a cost-effective method to integrate speech into a large language model (LLM), resulting in a Contextual Speech Model with Instruction-following/in-context-learning Capabilities (COSMIC) multi-modal LLM. Using GPT-3.5, we generate Speech Comprehension Test Question-Answer (SQA) pairs from speech transcriptions for supervised instruction tuning. With under 30 million trainable parameters and only 450 hours of English speech data, COSMIC demonstrates emerging capabilities in instruction-following and in-context learning. Equipped with such capabilities, COSMIC achieves a maximum 33.18 BLEU score in 0-shot EN-to-X speech to text translation (S2TT) and a significant boost in the 1-shot setting. Additionally, there is an average 25.8\% relative Word Error Rate (WER) reduction for 1-shot cross-domain adaptation. COSMIC exhibits a significant automatic speech recognition (ASR) accuracy gain in contextual biasing tasks due to its instruction-following capability.

cs.CL↗

arXiv2023

Improving Stability in Simultaneous Speech Translation: A Revision-Controllable Decoding Approach

Simultaneous Speech-to-Text translation serves a critical role in real-time crosslingual communication. Despite the advancements in recent years, challenges remain in achieving stability in the translation process, a concern primarily manifested in the flickering of partial results. In this paper, we propose a novel revision-controllable method designed to address this issue. Our method introduces an allowed revision window within the beam search pruning process to screen out candidate translations likely to cause extensive revisions, leading to a substantial reduction in flickering and, crucially, providing the capability to completely eliminate flickering. The experiments demonstrate the proposed method can significantly improve the decoding stability without compromising substantially on the translation quality.

cs.CL↗

arXiv2023

Frequency-astigmatism asymmetric nonlinear conversion of structured light lasers

Nonlinear optics of structured light has recently delivered intriguing fundamental physical phenomena in light-matter interactions and advanced applications from classical imaging to quantum informatics. The mutual interaction between spin, orbital angular momentum (OAM) and wavelength is extensively studied in such cases. In this work, we go beyond only considering OAM and wavelength by taking the nonlinear frequency conversion and transverse mode astigmatism conversion as two building blocks and investigating how single modes and complicated multiplexed modes evolve after them. In particular, We found a generalized law of nonlinear conversion structured light from experiments and theories, that the converted modes are highly related to the sequence of these two blocks, obeying an inherent (non)commutative rule in which. This effect not only creates extended structured laser modes but serve as new rules in nonlinear structured light manipulation.

physics.optics↗

arXiv2022

E-Branchformer: Branchformer with Enhanced merging for speech recognition

Conformer, combining convolution and self-attention sequentially to capture both local and global information, has shown remarkable performance and is currently regarded as the state-of-the-art for automatic speech recognition (ASR). Several other studies have explored integrating convolution and self-attention but they have not managed to match Conformer's performance. The recently introduced Branchformer achieves comparable performance to Conformer by using dedicated branches of convolution and self-attention and merging local and global context from each branch. In this paper, we propose E-Branchformer, which enhances Branchformer by applying an effective merging method and stacking additional point-wise modules. E-Branchformer sets new state-of-the-art word error rates (WERs) 1.81% and 3.65% on LibriSpeech test-clean and test-other sets without using any external training data.

eess.AS↗

arXiv2022

3D inhomogeneous self-accelerating beams

We propose and generate a new class of structured light fulfilling quantum-like coherent states based on a set of circular Airy vortex modes. Such coherent-state wave packets possess strong focus with both radial and angular self-accelerations, which exploit more general 3D inhomogeneous velocity control with global spatial symmetry of multilayer rotation akin to galactic kinematics, as termed galaxy waves. Galaxy waves are endowed with new degrees of freedom to control strong focusing and acceleration of 3D structured light, promising numerous applications in optical trapping, manufacturing, and nonlinear optics.

physics.optics↗

arXiv2021

SRU++: Pioneering Fast Recurrence with Attention for Speech Recognition

The Transformer architecture has been well adopted as a dominant architecture in most sequence transduction tasks including automatic speech recognition (ASR), since its attention mechanism excels in capturing long-range dependencies. While models built solely upon attention can be better parallelized than regular RNN, a novel network architecture, SRU++, was recently proposed. By combining the fast recurrence and attention mechanism, SRU++ exhibits strong capability in sequence modeling and achieves near-state-of-the-art results in various language modeling and machine translation tasks with improved compute efficiency. In this work, we present the advantages of applying SRU++ in ASR tasks by comparing with Conformer across multiple ASR benchmarks and study how the benefits can be generalized to long-form speech inputs. On the popular LibriSpeech benchmark, our SRU++ model achieves 2.0% / 4.7% WER on test-clean / test-other, showing competitive performances compared with the state-of-the-art Conformer encoder under the same set-up. Specifically, SRU++ can surpass Conformer on long-form speech input with a large margin, based on our analysis.

eess.AS↗

arXiv2021

Sensoring and Application of Multimodal Data for the Detection of Freezing of Gait in Parkinson's Disease

The accurate and reliable detection or prediction of freezing of gaits (FOG) is important for fall prevention in Parkinson's Disease (PD) and studying the physiological transitions during the occurrence of FOG. Integrating both commercial and self-designed sensors, a protocal has been designed to acquire multimodal physical and physiological information during FOG, including gait acceleration (ACC), electroencephalogram (EEG), electromyogram (EMG), and skin conductance (SC). Two tasks were designed to trigger FOG, including gait initiation failure and FOG during walking. A total number of 12 PD patients completed the experiments and produced a total length of 3 hours and 42 minutes of valid data. The FOG episodes were labeled by two qualified physicians. Each unimodal data and combinations have been used to detect FOG. Results showed that multimodal data benefit the detection of FOG. Among unimodal data, EEG had better discriminative ability than ACC and EMG. However, the acquisition of EEG are more complicated. Multimodal motional and electrophysiological data can also be used to study the physiological transition process during the occurrence of FOG and provide personalised interventions.

eess.SP↗