Modeling long range dependencies in sequential data is a fundamental step towards attaining human-level performance in many modalities such as text, vision, audio and video. While attention-based models are a popular and effective choice in modeling short-range interactions, their performance on tasks requiring long range reasoning has been largely inadequate. In an exciting result, Gu et al. (ICLR 2022) proposed the $\textit{Structured State Space}$ (S4) architecture delivering large gains over state-of-the-art models on several long-range tasks across various modalities. The core proposition of S4 is the parameterization of state matrices via a diagonal plus low rank structure, allowing efficient computation. In this work, we show that one can match the performance of S4 even without the low rank correction and thus assuming the state matrices to be diagonal. Our $\textit{Diagonal State Space}$ (DSS) model matches the performance of S4 on Long Range Arena tasks, speech classification on Speech Commands dataset, while being conceptually simpler and straightforward to implement.

We propose a parameter server system for distributed ML, which follows a Stale Synchronous Parallel (SSP) model of computation that maximizes the time computational workers spend doing useful work on ML algorithms, while still providing correctness guarantees. The parameter server provides an easy-to-use shared interface for read/write access to an ML model's values (parameters and variables), and the SSP model allows distributed workers to read older, stale versions of these values from a local cache, instead of waiting to get them from a central storage. This significantly increases the proportion of time workers spend computing, as opposed to waiting. Furthermore, the SSP model ensures ML algorithm correctness by limiting the maximum age of the stale values. We provide a proof of correctness under SSP, as well as empirical results demonstrating that the SSP model achieves faster algorithm convergence on several different ML problems, compared to fully-synchronous and asynchronous schemes.

Time series forecasting is crucial for applications across multiple domains and various scenarios. Although Transformers have dramatically advanced the landscape of forecasting, their effectiveness remains debated. Recent findings have indicated that simpler linear models might outperform complex Transformer-based approaches, highlighting the potential for more streamlined architectures. In this paper, we shift the focus from evaluating the overall Transformer architecture to specifically examining the effectiveness of self-attention for time series forecasting. To this end, we introduce a new architecture, Cross-Attention-only Time Series transformer (CATS), that rethinks the traditional transformer framework by eliminating self-attention and leveraging cross-attention mechanisms instead.

A partial label (PL) specifies a set of candidate labels for an instance and partial-label learning (PLL) trains multi-class classifiers with PLs.Recently, many methods that incorporate techniques from other domains have shown strong potential.The expectation that stronger techniques would enhance performance has resulted in prominent PLL methods becoming not only highly complicated but also quite different from one another, making it challenging to choose the best direction for future algorithm design.While it is exciting to see higher performance, this leaves open a fundamental question: what makes a PLL method effective?We present a comprehensive empirical analysis of this question and summarize the success of PLL so far into some minimal algorithm design principles.Our findings reveal that high accuracy on benchmark-simulated datasets with PLs can misleadingly amplify the perceived effectiveness of some general techniques, which may improve representation learning but have limited impact on addressing the inherent challenges of PLs. We further identify the common behavior among successful PLL methods as a progressive transition from uniform to one-hot pseudo-labels, highlighting the critical role of mini-batch PL purification in achieving top performance.Based on our findings, we introduce a minimal working algorithm that is surprisingly simple yet effective, and propose an improved strategy to implement the design principles, suggesting a promising direction for improvements in PLL.

Due to the recurrent structure of RNN, the long information propagation path poses limitations in capturing long-term dependencies, gradient explosion/vanishing issues, and inefficient sequential execution. Based on this, we propose a novel paradigm called Parallel Gated Network (PGN) as the new successor to RNN. PGN directly captures information from previous time steps through the designed Historical Information Extraction (HIE) layer and leverages gated mechanisms to select and fuse it with the current time step information. This reduces the information propagation path to \mathcal{O}(1), effectively addressing the limitations of RNN. To enhance PGN's performance in long-range time series forecasting tasks, we propose a novel temporal modeling framework called Temporal PGN (TPGN).

After many years of deliberating about the potential negative consequences of industrial automation for blue-collar workers, it is now clear that robots will also be coming for professionals. This may not be possible immediately. It seems certain that automation will allow organizations to automate many of the processes and tasks currently performed by humans in the coming years. The latest version of OpenAI's ChatGPT is causing a lot of concern. Its remarkable language generation abilities have sparked a legit AI arms race between tech giants such as Meta, Google, Apple, and Meta.

There is no better time to begin using artificial intelligence (AI) to optimize your marketing efforts. You can use AI to make marketing more effective. In fact, this could be one of the most disruptive technologies of our time. The key to becoming a marketing leader lies in the ability to identify what works best for your particular niche or industry and apply the latest in AI technology to optimize those efforts. AI can help you improve the effectiveness of your marketing efforts by helping you to automate, scale, and personalize them.

Dialog between drivers and speech-based vehicle interfaces can be used as an instrument to find out what drivers might be concerned, confused or curious about in driving simulator studies. Eliciting on-going conversation with drivers about topics that go beyond navigation, control of entertainment systems, or other traditional driving related tasks is important to getting drivers to engage with the activity in an open-ended fashion. In a structured improvisational Wizard of Oz study that took place in a highly immersive driving simulator, we engaged participant drivers (N=6) in an autonomous driving course where the vehicle spoke to drivers using computer-generated natural language speech. Using microanalyses of the drivers’ responses to the car’s utter- ances, we identify a set of topics that are expected and treated as appropriate by the participants in our study, as well as a set of topics and conversational strategies that are treated as inappropriate. We also show that it is just these unexpected, inappropriate utterances that eventually increase users’ trust in the system, make them more at ease, and raise the system’s acceptability as a communication partner.

We present in this paper an innovative solution to the challenge of building effective educational technologies that offer tailored instruction to each individual learner. The proposed solution in the form of a conversational intelligent tutoring system, called DeepTutor, has been developed as a web application that is accessible 24/7 through a browser from any device connected to the Internet. The success of several large scale experiments with high-school students using DeepTutor is a solid proof that conversational intelligent tutoring at scale over the web is possible.

In this work we consider a setting where we have a very large number of related tasks with few examples from each individual task. Rather than either learning each task individually (and having a large generalization error) or learning all the tasks together using a single hypothesis (and suffering a potentially large inherent error), we consider learning a small pool of {\em shared hypotheses}. Each task is then mapped to a single hypothesis in the pool (hard association). We derive VC dimension generalization bounds for our model, based on the number of tasks, shared hypothesis and the VC dimension of the hypotheses class. We conducted experiments with both synthetic problems and sentiment of reviews, which strongly support our approach.