Solving partial differential equations (PDEs) with neural networks (NNs) has shown great potential in various scientific and engineering fields. However, most existing NN solvers mainly focus on satisfying the given PDE formulas in the strong or weak sense, without explicitly considering some intrinsic physical properties, such as mass and momentum conservation, or energy dissipation. This limitation may result in nonphysical or unstable numerical solutions, particularly in long-term simulations. To address this issue, we propose ``Sidecar'', a novel framework that enhances the physical consistency of existing NN solvers for solving parabolic PDEs. Inspired by the time-dependent spectral renormalization approach, our Sidecar framework introduces a small network as a copilot, guiding the primary function-learning NN solver to respect the structure-preserving properties. Our framework is highly flexible, allowing the preservation of various physical quantities for different PDEs to be incorporated into a wide range of NN solvers. Experimental results on some benchmark problems demonstrate significant improvements brought by the proposed framework to both accuracy and structure preservation of existing NN solvers.

Federated Learning (FL) typically involves a large-scale, distributed system with individual user devices/servers training models locally and then aggregating their model updates on a trusted central server. Existing systems for FL often use an always-on server for model aggregation, which can be inefficient in terms of resource utilization. They may also be inelastic in their resource management. This is particularly exacerbated when aggregating model updates at scale in a highly dynamic environment with varying numbers of heterogeneous user devices/servers. We present LIFL, a lightweight and elastic serverless cloud platform with fine-grained resource management for efficient FL aggregation at scale. LIFL is enhanced by a streamlined, event-driven serverless design that eliminates the individual heavy-weight message broker and replaces inefficient container-based sidecars with lightweight eBPF-based proxies. We leverage shared memory processing to achieve high-performance communication for hierarchical aggregation, which is commonly adopted to speed up FL aggregation at scale. We further introduce locality-aware placement in LIFL to maximize the benefits of shared memory processing. LIFL precisely scales and carefully reuses the resources for hierarchical aggregation to achieve the highest degree of parallelism while minimizing the aggregation time and resource consumption. Our experimental results show that LIFL achieves significant improvement in resource efficiency and aggregation speed for supporting FL at scale, compared to existing serverful and serverless FL systems.

Multi-talker overlapped speech poses a significant challenge for speech recognition and diarization. Recent research indicated that these two tasks are inter-dependent and complementary, motivating us to explore a unified modeling method to address them in the context of overlapped speech. A recent study proposed a cost-effective method to convert a single-talker automatic speech recognition (ASR) system into a multi-talker one, by inserting a Sidecar separator into the frozen well-trained ASR model. Extending on this, we incorporate a diarization branch into the Sidecar, allowing for unified modeling of both ASR and diarization with a negligible overhead of only 768 parameters. The proposed method yields better ASR results compared to the baseline on LibriMix and LibriSpeechMix datasets. Moreover, without sophisticated customization on the diarization task, our method achieves acceptable diarization results on the two-speaker subset of CALLHOME with only a few adaptation steps.

Although automatic speech recognition (ASR) can perform well in common non-overlapping environments, sustaining performance in multi-talker overlapping speech recognition remains challenging. Recent research revealed that ASR model's encoder captures different levels of information with different layers -- the lower layers tend to have more acoustic information, and the upper layers more linguistic. This inspires us to develop a Sidecar separator to empower a well-trained ASR model for multi-talker scenarios by separating the mixed speech embedding between two suitable layers. We experimented with a wav2vec 2.0-based ASR model with a Sidecar mounted. By freezing the parameters of the original model and training only the Sidecar (8.7 M, 8.4% of all parameters), the proposed approach outperforms the previous state-of-the-art by a large margin for the 2-speaker mixed LibriMix dataset, reaching a word error rate (WER) of 10.36%; and obtains comparable results (7.56%) for LibriSpeechMix dataset when limited training.

On this episode of Recode Decode, hosted by Kara Swisher, Kara sits downs with Sunil Paul, the co-founder of Sidecar who recently penned a popular post for this site, "The scooter wars will be a bloodbath, and Uber will win." In this podcast, he elaborates on why that is and shares his thoughts about the broader transportation industry, including self-driving cars, bike-sharing and vertical lift and take-off vehicles like Larry Page's Kitty Hawk "flying car." Now primarily an investor, Paul also talks about why Sidecar couldn't compete with Uber and Lyft -- even though it created ride-hailing features that are now popular parts of their products. You can listen to Recode Decode on Apple Podcasts, Spotify, Pocket Casts, Overcast or wherever you listen to podcasts. Below, we've shared a lightly edited transcript of Kara's full conversation with Paul. Sunil Paul: Great to be here. Let's do a little background. You and I have known each other for ... A dog's age, as they say, like since D.C., in the early 90s. Can you explain how you were lucky enough to meet me then? Well, I think I first met you when I was AOL's internet product manager. And then, I started a company. I think you were the demo boy. That's what I think you were, weren't you? You showed me some demos. I think I was a demo boy. I recall demo boy-ing for Steve Case. You were working at AOL. What products did you work on there? How did you get there? What were you doing in D.C.? I came to D.C. to work on a space station. My early career I was an engineer. I helped do the early design for a space station. Then I got really interested in policy because Congress kept mucking around with the space station. I went, spent several years as a policy analyst on the Hill for [the] Office of Technology Assessment. While I was there, I started mucking around with this new thing, that was the early '90s, I started mucking around with this new thing called the internet and that ... Why did you pick AOL?

On this episode of Recode Decode, hosted by Kara Swisher, Kara sits downs with Sunil Paul, the co-founder of Sidecar who recently penned a popular post for this site, "The scooter wars will be a bloodbath, and Uber will win." In this podcast, he elaborates on why that is and shares his thoughts about the broader transportation industry, including self-driving cars, bike-sharing and vertical lift and take-off vehicles like Larry Page's Kitty Hawk "flying car." Now primarily an investor, Paul also talks about why Sidecar couldn't compete with Uber and Lyft -- even though it created ride-hailing features that are now popular parts of their products. You can listen to Recode Decode on Apple Podcasts, Spotify, Pocket Casts, Overcast or wherever you listen to podcasts. Below, we've shared a lightly edited transcript of Kara's full conversation with Paul. Sunil Paul: Great to be here. Let's do a little background. You and I have known each other for ... A dog's age, as they say, like since D.C., in the early 90s. Can you explain how you were lucky enough to meet me then? Well, I think I first met you when I was AOL's internet product manager. And then, I started a company. I think you were the demo boy. That's what I think you were, weren't you? You showed me some demos. I think I was a demo boy. I recall demo boy-ing for Steve Case. You were working at AOL. What products did you work on there? How did you get there? What were you doing in D.C.? I came to D.C. to work on a space station. My early career I was an engineer. I helped do the early design for a space station. Then I got really interested in policy because Congress kept mucking around with the space station. I went, spent several years as a policy analyst on the Hill for [the] Office of Technology Assessment. While I was there, I started mucking around with this new thing, that was the early '90s, I started mucking around with this new thing called the internet and that ... Why did you pick AOL?

Machine learning underpins Sidecar's optimization technology for product listing ads, and we're not shy about making that fact known. But we understand that the term might sound a little bit sci-fi to some. However, since it's so central to what we do and who we are, we thought a simple explanation of machine learning was in order -- no computer science degree required. At its simplest, machine learning means giving computers the power to teach themselves to make decisions using historical examples, rather than explicitly programming them to perform a task. To illustrate this concept, I'm going to borrow an example from The Data Skeptic, a popular -- and highly recommended -- podcast by Kyle Polich.