Search

We answer in the negative Siegel's problem for $G$-functions, as formulated by Fischler and Rivoal. Roughly, we prove that there are $G$-functions that cannot be written as polynomial expressions in algebraic pullbacks of hypergeometric functions; our examples satisfy differential equations of order two, which is the smallest possible. In fact, we construct infinitely many non-equivalent rank-two local systems of geometric origin which are not algebraic pullbacks of hypergeometric local systems, thereby providing further counterexamples to Dwork's conjecture and answering a question by Krammer. The main ingredients of the proof are a Lie algebra version of Goursat's lemma, the monodromy computations of hypergeometric local systems due to Beukers and Heckman, as well as results on invariant trace fields of Fuchsian groups., Comment: 24 pages--comments welcome!

We formulate a conjecture classifying algebraic solutions to (possibly non-linear) algebraic differential equations, in terms of the primes appearing in the denominators of the coefficients of their Taylor expansion at a non-singular point. For linear differential equations, this conjecture is a strengthening of the Grothendieck-Katz $p$-curvature conjecture. We prove the conjecture for many differential equations and initial conditions of algebro-geometric interest. For linear differential equations, we prove it for Picard-Fuchs equations at initial conditions corresponding to cycle classes, among other cases. For non-linear differential equations, we prove it for isomonodromy differential equations, such as the Painlev\'e VI equation and Schlesinger system, at initial conditions corresponding to Picard-Fuchs equations. We draw a number of algebro-geometric consequences from the proofs.

ELIZA, created by Joseph Weizenbaum at MIT in the early 1960s, is usually considered the world's first chatbot. It was developed in MAD-SLIP on MIT's CTSS, the world's first time-sharing system, on an IBM 7094. We discovered an original ELIZA printout in Prof. Weizenbaum's archives at MIT, including an early version of the famous DOCTOR script, a nearly complete version of the MAD-SLIP code, and various support functions in MAD and FAP. Here we describe the reanimation of this original ELIZA on a restored CTSS, itself running on an emulated IBM 7094. The entire stack is open source, so that any user of a unix-like OS can run the world's first chatbot on the world's first time-sharing system., Comment: In review

Graph-based machine learning methods are useful tools in the identification and prediction of variation in genetic data. In particular, the comprehension of phenotypic effects at the cellular level is an accelerating research area in pharmacogenomics. In this article, a novel graph theoretic approach is proposed to infer a co-occurrence network from 16S microbiome data. The approach is specialised to handle datasets containing a small number of samples. Small datasets exacerbate the significant challenges faced by biological data, which exhibit properties such as sparsity, compositionality, and complexity of interactions. Methodologies are also proposed to enrich and statistically filter the inferred networks. The utility of the proposed method lies in that it extracts an informative network from small sampled data that is not only feature-rich, but also biologically meaningful and statistically significant. Although specialised for small data sets, which are abundant, it can be generally applied to any small-sampled dataset, and can also be extended to integrate multi-omics data. The proposed methodology is tested on a data set of chickens vaccinated against and challenged by the protozoan parasite Eimeria tenella. The raw genetic reads are processed, and networks inferred to describe the ecosystems of the chicken intestines under three different stages of disease progression. Analysis of the expression of network features derive biologically intuitive conclusions from purely statistical methods. For example, there is a clear evolution in the distribution of node features in line with the progression of the disease. The distributions also reveal clusters of species interacting mutualistically and parasitically, as expected. Moreover, a specific sub-network is found to persist through all experimental conditions, representative of a persistent microbiome., Comment: 12 pages, 7 figures, under review for a special issue of ACM/IEEE TCBB journal

The use of machine learning and AI on electronic health records (EHRs) holds substantial potential for clinical insight. However, this approach faces challenges due to data heterogeneity, sparsity, temporal misalignment, and limited labeled outcomes. In this context, we leverage a linked EHR dataset of approximately one million de-identified individuals from Bristol, North Somerset, and South Gloucestershire, UK, to characterize urinary tract infections (UTIs). We implemented a data pre-processing and curation pipeline that transforms the raw EHR data into a structured format suitable for developing predictive models focused on data fairness, accountability and transparency. Given the limited availability and biases of ground truth UTI outcomes, we introduce a UTI risk estimation framework informed by clinical expertise to estimate UTI risk across individual patient timelines. Pairwise XGBoost models are trained using this framework to differentiate UTI risk categories with explainable AI techniques applied to identify key predictors and support interpretability. Our findings reveal differences in clinical and demographic predictors across risk groups. While this study highlights the potential of AI-driven insights to support UTI clinical decision-making, further investigation of patient sub-strata and extensive validation are needed to ensure robustness and applicability in clinical practice.

Modern transformer-based deep neural networks present unique technical challenges for effective acceleration in real-world applications. Apart from the vast amount of linear operations needed due to their sizes, modern transformer models are increasingly reliance on precise non-linear computations that make traditional low-bitwidth quantization methods and fixed-dataflow matrix accelerators ineffective for end-to-end acceleration. To address this need to accelerate both linear and non-linear operations in a unified and programmable framework, this paper introduces TATAA. TATAA employs 8-bit integer (int8) arithmetic for quantized linear layer operations through post-training quantization, while it relies on bfloat16 floating-point arithmetic to approximate non-linear layers of a transformer model. TATAA hardware features a transformable arithmetic architecture that supports both formats during runtime with minimal overhead, enabling it to switch between a systolic array mode for int8 matrix multiplications and a SIMD mode for vectorized bfloat16 operations. An end-to-end compiler is presented to enable flexible mapping from emerging transformer models to the proposed hardware. Experimental results indicate that our mixed-precision design incurs only 0.14% to 1.16% accuracy drop when compared with the pre-trained single-precision transformer models across a range of vision, language, and generative text applications. Our prototype implementation on the Alveo U280 FPGA currently achieves 2935.2 GOPS throughput on linear layers and a maximum of 189.5 GFLOPS for non-linear operations, outperforming related works by up to 1.45x in end-to-end throughput and 2.29x in DSP efficiency, while achieving 2.19x higher power efficiency than modern NVIDIA RTX4090 GPU.

Attention is the cornerstone of modern Large Language Models (LLMs). Yet its quadratic complexity hinders efficiency and scalability, especially for long-context processing. A promising approach is to leverage sparsity in attention. However, existing sparsity-based solutions predominantly rely on predefined patterns or heuristics at the attention head level, struggling to adapt dynamically to different contexts efficiently. We propose SeerAttention, a simple yet effective attention mechanism that directly learns the block-level attention sparsity from the LLM itself. Inspired by the gating mechanism in Mixture of Experts (MoE), SeerAttention augments the conventional attention with a learnable gate that selectively activates important blocks within the attention map. Specifically, the gate first pools the query (Q) and key (K) tensors along the sequence dimension and processes them through learnable linear layers. The resulting matrices are then multiplied together to produce the gating scores, which are used to predict block-level attention sparsity. Combined with our block-sparse FlashAttention kernel, SeerAttention can achieve significant speedup on GPUs. When applied to pre-trained LLMs, SeerAttention only requires training the gate parameters in a lightweight self-distillation manner, allowing rapid convergence. Our evaluation results demonstrate that SeerAttention achieves better model accuracy and lower latency for long-context pre-filling compared to prior methods. Code is available at: https://github.com/microsoft/SeerAttention

For each smooth curve over a finite field, after puncturing it at finitely many points, we construct local systems on it of geometric origin which do not come from a family of abelian varieties. We do so by proving a criterion which must be satisfied by local systems which do come from abelian varieties, inspired by an analogous Hodge theoretic criterion in characteristic zero.

We present foundation language models developed to power Apple Intelligence features, including a ~3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks efficiently, accurately, and responsibly. This report describes the model architecture, the data used to train the model, the training process, how the models are optimized for inference, and the evaluation results. We highlight our focus on Responsible AI and how the principles are applied throughout the model development.

The process of human speech production involves coordinated respiratory action to elicit acoustic speech signals. Typically, speech is produced when air is forced from the lungs and is modulated by the vocal tract, where such actions are interspersed by moments of breathing in air (inhalation) to refill the lungs again. Respiratory rate (RR) is a vital metric that is used to assess the overall health, fitness, and general well-being of an individual. Existing approaches to measure RR (number of breaths one takes in a minute) are performed using specialized equipment or training. Studies have demonstrated that machine learning algorithms can be used to estimate RR using bio-sensor signals as input. Speech-based estimation of RR can offer an effective approach to measure the vital metric without requiring any specialized equipment or sensors. This work investigates a machine learning based approach to estimate RR from speech segments obtained from subjects speaking to a close-talking microphone device. Data were collected from N=26 individuals, where the groundtruth RR was obtained through commercial grade chest-belts and then manually corrected for any errors. A convolutional long-short term memory network (Conv-LSTM) is proposed to estimate respiration time-series data from the speech signal. We demonstrate that the use of pre-trained representations obtained from a foundation model, such as Wav2Vec2, can be used to estimate respiration-time-series with low root-mean-squared error and high correlation coefficient, when compared with the baseline. The model-driven time series can be used to estimate $RR$ with a low mean absolute error (MAE) ~ 1.6 breaths/min., Comment: 8 pages, 6 figures, BioKDD workshop paper

With 95% of Internet traffic now encrypted, an effective approach to classifying this traffic is crucial for network security and management. This paper introduces ECHO -- a novel optimization process for ML/DL-based encrypted traffic classification. ECHO targets both classification time and memory utilization and incorporates two innovative techniques. The first component, HO (Hyperparameter Optimization of binnings), aims at creating efficient traffic representations. While previous research often uses representations that map packet sizes and packet arrival times to fixed-sized bins, we show that non-uniform binnings are significantly more efficient. These non-uniform binnings are derived by employing a hyperparameter optimization algorithm in the training stage. HO significantly improves accuracy given a required representation size, or, equivalently, achieves comparable accuracy using smaller representations. Then, we introduce EC (Early Classification of traffic), which enables faster classification using a cascade of classifiers adapted for different exit times, where classification is based on the level of confidence. EC reduces the average classification latency by up to 90\%. Remarkably, this method not only maintains classification accuracy but also, in certain cases, improves it. Using three publicly available datasets, we demonstrate that the combined method, Early Classification with Hyperparameter Optimization (ECHO), leads to a significant improvement in classification efficiency.

This paper addresses the problem of accurately estimating a function on one domain when only its discrete samples are available on another domain. To answer this challenge, we utilize a neural network, which we train to incorporate prior knowledge of the function. In addition, by carefully analyzing the problem, we obtain a bound on the error over the extrapolation domain and define a condition number for this problem that quantifies the level of difficulty of the setup. Compared to other machine learning methods that provide time series prediction, such as transformers, our approach is suitable for setups where the interpolation and extrapolation regions are general subdomains and, in particular, manifolds. In addition, our construction leads to an improved loss function that helps us boost the accuracy and robustness of our neural network. We conduct comprehensive numerical tests and comparisons of our extrapolation versus standard methods. The results illustrate the effectiveness of our approach in various scenarios., Comment: 32 pages, 11 figures

Eye-tracking technology is integral to numerous consumer electronics applications, particularly in the realm of virtual and augmented reality (VR/AR). These applications demand solutions that excel in three crucial aspects: low-latency, low-power consumption, and precision. Yet, achieving optimal performance across all these fronts presents a formidable challenge, necessitating a balance between sophisticated algorithms and efficient backend hardware implementations. In this study, we tackle this challenge through a synergistic software/hardware co-design of the system with an event camera. Leveraging the inherent sparsity of event-based input data, we integrate a novel sparse FPGA dataflow accelerator customized for submanifold sparse convolution neural networks (SCNN). The SCNN implemented on the accelerator can efficiently extract the embedding feature vector from each representation of event slices by only processing the non-zero activations. Subsequently, these vectors undergo further processing by a gated recurrent unit (GRU) and a fully connected layer on the host CPU to generate the eye centers. Deployment and evaluation of our system reveal outstanding performance metrics. On the Event-based Eye-Tracking-AIS2024 dataset, our system achieves 81% p5 accuracy, 99.5% p10 accuracy, and 3.71 Mean Euclidean Distance with 0.7 ms latency while only consuming 2.29 mJ per inference. Notably, our solution opens up opportunities for future eye-tracking systems. Code is available at https://github.com/CASR-HKU/ESDA/tree/eye_tracking., Comment: Accepted to CVPR 2024 workshop, AIS: Vision, Graphics, and AI for Streaming

This survey reviews the AIS 2024 Event-Based Eye Tracking (EET) Challenge. The task of the challenge focuses on processing eye movement recorded with event cameras and predicting the pupil center of the eye. The challenge emphasizes efficient eye tracking with event cameras to achieve good task accuracy and efficiency trade-off. During the challenge period, 38 participants registered for the Kaggle competition, and 8 teams submitted a challenge factsheet. The novel and diverse methods from the submitted factsheets are reviewed and analyzed in this survey to advance future event-based eye tracking research., Comment: Qinyu Chen is the corresponding author