Search

The rapid proliferation of generative AI has raised questions about the competitiveness of lower-parameter, locally tunable, open-weight models relative to high-parameter, API-guarded, closed-weight models in terms of performance, domain adaptation, cost, and generalization. Centering under-resourced yet risk-intolerant settings in government, research, and healthcare, we see for-profit closed-weight models as incompatible with requirements for transparency, privacy, adaptability, and standards of evidence. Yet the performance penalty in using open-weight models, especially in low-data and low-resource settings, is unclear. We assess the feasibility of using smaller, open-weight models to replace GPT-4-Turbo in zero-shot, few-shot, and fine-tuned regimes, assuming access to only a single, low-cost GPU. We assess value-sensitive issues around bias, privacy, and abstention on three additional tasks relevant to those topics. We find that with relatively low effort, very low absolute monetary cost, and relatively little data for fine-tuning, small open-weight models can achieve competitive performance in domain-adapted tasks without sacrificing generality. We then run experiments considering practical issues in bias, privacy, and hallucination risk, finding that open models offer several benefits over closed models. We intend this work as a case study in understanding the opportunity cost of reproducibility and transparency over for-profit state-of-the-art zero shot performance, finding this cost to be marginal under realistic settings., Comment: Accepted at the ACM Conference on Fairness, Accountability, and Transparency (FAccT) 2024

The emerging CXL.mem standard provides a new type of byte-addressable remote memory with a variety of memory types and hierarchies. With CXL.mem, multiple layers of memory -- e.g., local DRAM and CXL-attached remote memory at different locations -- are exposed to operating systems and user applications, bringing new challenges and research opportunities. Unfortunately, since CXL.mem devices are not commercially available, it is difficult for researchers to conduct systems research that uses CXL.mem. In this paper, we present our ongoing work, CXLMemSim, a fast and lightweight CXL.mem simulator for performance characterization. CXLMemSim uses a performance model driven using performance monitoring events, which are supported by most commodity processors. Specifically, CXLMemSim attaches to an existing, unmodified program, and divides the execution of the program into multiple epochs; once an epoch finishes, CXLMemSim collects performance monitoring events and calculates the simulated execution time of the epoch based on these events. Through this method, CXLMemSim avoids the performance overhead of a full-system simulator (e.g., Gem5) and allows the memory hierarchy and latency to be easily adjusted, enabling research such as memory scheduling for complex applications. Our preliminary evaluation shows that CXLMemSim slows down the execution of the attached program by 4.41x on average for real-world applications.

The ability to modify and extend an operating system is an important feature for improving a system's security, reliability, and performance. The extended Berkeley Packet Filters (eBPF) ecosystem has emerged as the standard mechanism for extending the Linux kernel and has recently been ported to Windows. eBPF programs inject new logic into the kernel that the system will execute before or after existing logic. While the eBPF ecosystem provides a flexible mechanism for kernel extension, it is difficult for developers to write eBPF programs today. An eBPF developer must have deep knowledge of the internals of the operating system to determine where to place logic and cope with programming limitations on the control flow and data accesses of their eBPF program enforced by the eBPF verifier. This paper presents KEN, an alternative framework that alleviates the difficulty of writing an eBPF program by allowing Kernel Extensions to be written in Natural language. KEN uses recent advances in large language models (LLMs) to synthesize an eBPF program given a user's English language prompt. To ensure that LLM's output is semantically equivalent to the user's prompt, KEN employs a combination of LLM-empowered program comprehension, symbolic execution, and a series of feedback loops. KEN's key novelty is the combination of these techniques. In particular, the system uses symbolic execution in a novel structure that allows it to combine the results of program synthesis and program comprehension and build on the recent success that LLMs have shown for each of these tasks individually. To evaluate KEN, we developed a new corpus of natural language prompts for eBPF programs. We show that KEN produces correct eBPF programs on 80% which is an improvement of a factor of 2.67 compared to an LLM-empowered program synthesis baseline.

In kernel-centric operations, the uprobe component of eBPF frequently encounters performance bottlenecks, largely attributed to the overheads borne by context switches. Transitioning eBPF operations to user space bypasses these hindrances, thereby optimizing performance. This also enhances configurability and obviates the necessity for root access or privileges for kernel eBPF, subsequently minimizing the kernel attack surface. This paper introduces bpftime, a novel user-space eBPF runtime, which leverages binary rewriting to implement uprobe and syscall hook capabilities. Through bpftime, userspace uprobes achieve a 10x speed enhancement compared to their kernel counterparts without requiring dual context switches. Additionally, this runtime facilitates the programmatic hooking of syscalls within a process, both safely and efficiently. Bpftime can be seamlessly attached to any running process, limiting the need for either a restart or manual recompilation. Our implementation also extends to interprocess eBPF Maps within shared memory, catering to summary aggregation or control plane communication requirements. Compatibility with existing eBPF toolchains such as clang and libbpf is maintained, not only simplifying the development of user-space eBPF without necessitating any modifications but also supporting CO-RE through BTF. Through bpftime, we not only enhance uprobe performance but also extend the versatility and user-friendliness of eBPF runtime in user space, paving the way for more efficient and secure kernel operations.

CXL has been the emerging technology for expanding memory for both the host CPU and device accelerators with load/store interface. Extending memory coherency to the PCIe root complex makes the codesign more flexible in that you can access the memory with coherency using your near-device computability. Since the capacity demand with tolerable latency and bandwidth is growing, we need to come up with a new hardware-software codesign way to offload the synthesized memory operations to the CXL endpoint, CXL switch or near CXL root complex cores like Intel DSA to fetch data; the CPU or accelerators can calculate other stuff in the backend. On CXL done loading, the data will be put into L1 if capacity fits, and the in-core ROB will be notified by mailbox and resume the calculation on the previous hardware context. Since the distance(timing window) of the load instruction sequence is unknown, a profiling-guided way of codegening and adaptively updating offloaded code will be required for a long-running job. We propose to evaluate CXLMemUring the modified BOOMv3 with added in-core-logic and CXL endpoint access simulation using CHI, and we will add a weaker RISCV Core near endpoint for code offloading, and the codegening will be based on program analysis with traditional profiling guided way.

Nine language-vision AI models trained on web scrapes with the Contrastive Language-Image Pretraining (CLIP) objective are evaluated for evidence of a bias studied by psychologists: the sexual objectification of girls and women, which occurs when a person's human characteristics, such as emotions, are disregarded and the person is treated as a body. We replicate three experiments in psychology quantifying sexual objectification and show that the phenomena persist in AI. A first experiment uses standardized images of women from the Sexual OBjectification and EMotion Database, and finds that human characteristics are disassociated from images of objectified women: the model's recognition of emotional state is mediated by whether the subject is fully or partially clothed. Embedding association tests (EATs) return significant effect sizes for both anger (d >0.80) and sadness (d >0.50), associating images of fully clothed subjects with emotions. GRAD-CAM saliency maps highlight that CLIP gets distracted from emotional expressions in objectified images. A second experiment measures the effect in a representative application: an automatic image captioner (Antarctic Captions) includes words denoting emotion less than 50% as often for images of partially clothed women than for images of fully clothed women. A third experiment finds that images of female professionals (scientists, doctors, executives) are likely to be associated with sexual descriptions relative to images of male professionals. A fourth experiment shows that a prompt of "a [age] year old girl" generates sexualized images (as determined by an NSFW classifier) up to 73% of the time for VQGAN-CLIP and Stable Diffusion; the corresponding rate for boys never surpasses 9%. The evidence indicates that language-vision AI models trained on web scrapes learn biases of sexual objectification, which propagate to downstream applications., Comment: 12 pages, 4 figures, 2 tables

Radiotherapy is a treatment where radiation is used to eliminate cancer cells. The delineation of organs-at-risk (OARs) is a vital step in radiotherapy treatment planning to avoid damage to healthy organs. For nasopharyngeal cancer, more than 20 OARs are needed to be precisely segmented in advance. The challenge of this task lies in complex anatomical structure, low-contrast organ contours, and the extremely imbalanced size between large and small organs. Common segmentation methods that treat them equally would generally lead to inaccurate small-organ labeling. We propose a novel two-stage deep neural network, FocusNetv2, to solve this challenging problem by automatically locating, ROI-pooling, and segmenting small organs with specifically designed small-organ localization and segmentation sub-networks while maintaining the accuracy of large organ segmentation. In addition to our original FocusNet, we employ a novel adversarial shape constraint on small organs to ensure the consistency between estimated small-organ shapes and organ shape prior knowledge. Our proposed framework is extensively tested on both self-collected dataset of 1,164 CT scans and the MICCAI Head and Neck Auto Segmentation Challenge 2015 dataset, which shows superior performance compared with state-of-the-art head and neck OAR segmentation methods., Comment: Accepted by Medical Image Analysis

As a new programming paradigm, deep learning has expanded its application to many real-world problems. At the same time, deep learning based software are found to be vulnerable to adversarial attacks. Though various defense mechanisms have been proposed to improve robustness of deep learning software, many of them are ineffective against adaptive attacks. In this work, we propose a novel characterization to distinguish adversarial examples from benign ones based on the observation that adversarial examples are significantly less robust than benign ones. As existing robustness measurement does not scale to large networks, we propose a novel defense framework, named attack as defense (A2D), to detect adversarial examples by effectively evaluating an example's robustness. A2D uses the cost of attacking an input for robustness evaluation and identifies those less robust examples as adversarial since less robust examples are easier to attack. Extensive experiment results on MNIST, CIFAR10 and ImageNet show that A2D is more effective than recent promising approaches. We also evaluate our defence against potential adaptive attacks and show that A2D is effective in defending carefully designed adaptive attacks, e.g., the attack success rate drops to 0% on CIFAR10.

In this paper, we propose an end-to-end deep neural network for solving the problem of imbalanced large and small organ segmentation in head and neck (HaN) CT images. To conduct radiotherapy planning for nasopharyngeal cancer, more than 10 organs-at-risk (normal organs) need to be precisely segmented in advance. However, the size ratio between large and small organs in the head could reach hundreds. Directly using such imbalanced organ annotations to train deep neural networks generally leads to inaccurate small-organ label maps. We propose a novel end-to-end deep neural network to solve this challenging problem by automatically locating, ROI-pooling, and segmenting small organs with specifically designed small-organ sub-networks while maintaining the accuracy of large organ segmentation. A strong main network with densely connected atrous spatial pyramid pooling and squeeze-and-excitation modules is used for segmenting large organs, where large organs' label maps are directly output. For small organs, their probabilistic locations instead of label maps are estimated by the main network. High-resolution and multi-scale feature volumes for each small organ are ROI-pooled according to their locations and are fed into small-organ networks for accurate segmenting small organs. Our proposed network is extensively tested on both collected real data and the \emph{MICCAI Head and Neck Auto Segmentation Challenge 2015} dataset, and shows superior performance compared with state-of-the-art segmentation methods., Comment: MICCAI 2019

While the role of humans is increasingly recognized in machine learning community, representation of and interaction with models in current human-in-the-loop machine learning (HITL-ML) approaches are too low-level and far-removed from human's conceptual models. We demonstrate HEIDL, a prototype HITL-ML system that exposes the machine-learned model through high-level, explainable linguistic expressions formed of predicates representing semantic structure of text. In HEIDL, human's role is elevated from simply evaluating model predictions to interpreting and even updating the model logic directly by enabling interaction with rule predicates themselves. Raising the currency of interaction to such semantic levels calls for new interaction paradigms between humans and machines that result in improved productivity for text analytics model development process. Moreover, by involving humans in the process, the human-machine co-created models generalize better to unseen data as domain experts are able to instill their expertise by extrapolating from what has been learned by automated algorithms from few labelled data.

The Back-n white neutron beam line, which uses back-streaming white neutrons from the spallation target of the China Spallation Neutron Source, is used for nuclear data measurements. A Micromegas-based neutron detector with two variants was specially developed to measure the beam spot distribution for this beam line. In this article, the design, fabrication, and characterization of the detector are described. The results of the detector performance tests are presented, which include the relative electron transparency, the gain and the gain uniformity, and the neutron beam profile reconstruction capability. The result of the first measurement of the Back-n neutron beam spot distribution is also presented., Comment: 31 pages, 18 figures

Thesis: S.M. in Engineering and Management, Massachusetts Institute of Technology, System Design and Management Program, 2017., Cataloged from PDF version of thesis., Includes bibliographical references (pages 116-121)., The popularization of virtual reality (VR) and augmented reality (AR) applications gives designers and developers tremendous opportunities to shape the evolution of new interaction paradigms. Current human machine interaction models are built on top of the command and control metaphors dating back to the industrial revolution, where input has to be initiated with user's full conscious thought in a non-parallel and discrete manner. Further, many people experience the adverse side effects of software products that are designed with the goal of increasing engagement. Mental states like email anxiety are driving down the wellness and productivity of the general population, and as VR and AR devices replace our physical spaces with digital experiences, these problems will become far more serious. This thesis explores methodologies for designing to promote the state of Flow through VR and AR experiences that bring interactions with virtual information closer to how humans consciously and subconsciously perceive, process, and interact with the real world. Designing interfaces to promote the state of Flow has several benefits. First, it requires interfaces to respect people's focus, which increases productivity and quality. Second, Flow has been shown to give people a sense of enjoyment, leading people to feel happier and more fulfilled. Drawing on research suggesting that encouraging the characteristic expressions of Flow such as focus and an implemental mindset can help induce the Flow state, the thesis explores techniques VR and AR applications can use to promote the state of Flow and overall wellbeing through the design of their interfaces., by Yiwei "Sophia" Yang., S.M. in Engineering and Management

To validate nuclear data and code in the neutronics design of a hybrid reactor with thorium, integral experiments in two kinds of benchmark thorium assemblies with a D-T fusion neutron source have been performed. The one kind of 1D assemblies consists of polyethylene and depleted uranium shells. The other kind of 2D assemblies consists of three thorium oxide cylinders. The capture reaction rates, fission reaction rates, and (n, 2n) reaction rates in 232Th in the assemblies are measured by ThO2 foils. The leakage neutron spectra from the ThO2 cylinders are measured by a liquid scintillation detector. The experimental uncertainties in all the results are analyzed. The measured results are compared to the calculated ones with MCNP code and ENDF/B-VII.0 library data.

To validate nuclear data and code in the neutronics design of a hybrid reactor with thorium, integral experiments in two kinds of benchmark thorium assemblies with a D-T fusion neutron source have been performed. The one kind of 1D assemblies consists of polyethylene and depleted uranium shells. The other kind of 2D assemblies consists of three thorium oxide cylinders. The capture reaction rates, fission reaction rates, and (n, 2n) reaction rates in 232Th in the assemblies are measured by ThO2 foils. The leakage neutron spectra from the ThO2 cylinders are measured by a liquid scintillation detector. The experimental uncertainties in all the results are analyzed. The measured results are compared to the calculated ones with MCNP code and ENDF/B-VII.0 library data.