A large class of non-smooth practical optimization problems can be written as minimization of a sum of smooth and partly smooth functions. We consider such structured problems which also depend on a parameter vector and study the problem of differentiating its solution mapping with respect to the parameter which has far reaching applications in sensitivity analysis and parameter learning optmization problems. We show that under partial smoothness and other mild assumptions, Automatic Differentiation (AD) of the sequence generated by proximal splitting algorithms converges to the derivative of the solution mapping. For a variant of automatic differentiation, which we call Fixed-Point Automatic Differentiation (FPAD), we remedy the memory overhead problem of the Reverse Mode AD and moreover provide faster convergence theoretically. We numerically illustrate the convergence and convergence rates of AD and FPAD on Lasso and Group Lasso problems and demonstrate the working of FPAD on prototypical practical image denoising problem by learning the regularization term.

Online bipartite matching is a fundamental problem in online algorithms. The goal is to match two sets of vertices to maximize the sum of the edge weights, where for one set of vertices, each vertex and its corresponding edge weights appear in a sequence. Currently, in the practical recommendation system or search engine, the weights are decided by the inner product between the deep representation of a user and the deep representation of an item. The standard online matching needs to pay $nd$ time to linear scan all the $n$ items, computing weight (assuming each representation vector has length $d$), and then decide the matching based on the weights. However, in reality, the $n$ could be very large, e.g. in online e-commerce platforms. Thus, improving the time of computing weights is a problem of practical significance. In this work, we provide the theoretical foundation for computing the weights approximately. We show that, with our proposed randomized data structures, the weights can be computed in sublinear time while still preserving the competitive ratio of the matching algorithm.

We present Branch-Train-Merge (BTM), a communication-efficient algorithm for embarrassingly parallel training of large language models (LLMs). We show it is possible to independently train subparts of a new class of LLMs on different subsets of the data, eliminating the massive multi-node synchronization currently required to train LLMs. BTM learns a set of independent expert LMs (ELMs), each specialized to a different textual domain, such as scientific or legal text. These ELMs can be added and removed to update data coverage, ensembled to generalize to new domains, or averaged to collapse back to a single LM for efficient inference. New ELMs are learned by branching from (mixtures of) ELMs in the current set, further training the parameters on data for the new domain, and then merging the resulting model back into the set for future use. Experiments show that BTM improves in- and out-of-domain perplexities as compared to GPT-style Transformer LMs, when controlling for training cost. Through extensive analysis, we show that these results are robust to different ELM initialization schemes, but require expert domain specialization; LM ensembles with random data splits do not perform well. We also present a study of scaling BTM into a new corpus of 64 domains (192B whitespace-separated tokens in total); the resulting LM (22.4B total parameters) performs as well as a Transformer LM trained with 2.5 times more compute. These gains grow with the number of domains, suggesting more aggressive parallelism could be used to efficiently train larger models in future work.

Type 2 Diabetes is one of the most major and fatal diseases known to human beings, where thousands of people are subjected to the onset of Type 2 Diabetes every year. However, the diagnosis and prevention of Type 2 Diabetes are relatively costly in today's scenario; hence, the use of machine learning and deep learning techniques is gaining momentum for predicting the onset of Type 2 Diabetes. This research aims to increase the accuracy and Area Under the Curve (AUC) metric while improving the processing time for predicting the onset of Type 2 Diabetes. The proposed system consists of a deep learning technique that uses the Support Vector Machine (SVM) algorithm along with the Radial Base Function (RBF) along with the Long Short-term Memory Layer (LSTM) for prediction of onset of Type 2 Diabetes. The proposed solution provides an average accuracy of 86.31 % and an average AUC value of 0.8270 or 82.70 %, with an improvement of 3.8 milliseconds in the processing. Radial Base Function (RBF) kernel and the LSTM layer enhance the prediction accuracy and AUC metric from the current industry standard, making it more feasible for practical use without compromising the processing time.

The logic of hybrid MKNF (minimal knowledge and negation as failure) is a powerful knowledge representation language that elegantly pairs ASP (answer set programming) with ontologies. Disjunctive rules are a desirable extension to normal rule-based reasoning and typically semantic frameworks designed for normal knowledge bases need substantial restructuring to support disjunctive rules. Alternatively, one may lift characterizations of normal rules to support disjunctive rules by inducing a collection of normal knowledge bases, each with the same body and a single atom in its head. In this work, we refer to a set of such normal knowledge bases as a head-cut of a disjunctive knowledge base. The question arises as to whether the semantics of disjunctive hybrid MKNF knowledge bases can be characterized using fixpoint constructions with head-cuts. Earlier, we have shown that head-cuts can be paired with fixpoint operators to capture the two-valued MKNF models of disjunctive hybrid MKNF knowledge bases. Three-valued semantics extends two-valued semantics with the ability to express partial information. In this work, we present a fixpoint construction that leverages head-cuts using an operator that iteratively captures three-valued models of hybrid MKNF knowledge bases with disjunctive rules. This characterization also captures partial stable models of disjunctive logic programs since a program can be expressed as a disjunctive hybrid MKNF knowledge base with an empty ontology. We elaborate on a relationship between this characterization and approximators in AFT (approximation fixpoint theory) for normal hybrid MKNF knowledge bases.

Although masked language models are highly performant and widely adopted by NLP practitioners, they can not be easily used for autoregressive language modelling (next word prediction and sequence probability estimation). We present an LSTM-based autoregressive language model which uses prefix embeddings (from a pretrained masked language model) via fusion (e.g. concatenation) to obtain a richer context representation for language modelling. We find that fusion helps reliably in lowering the perplexity (16.74 $\rightarrow$ 15.80), which is even preserved after a transfer to a dataset from a different domain than the training data. We also evaluate the best-performing fusion model by correlating its next word surprisal estimates with human reading times. Contradicting our expectation, and despite the improvement in perplexity overall, the correlation remains the same as for the baseline model. Lastly, while we focus on language models pre-trained on text as the sources for the fusion, our approach can be possibly extended to fuse any information represented as a fixed-size vector into an auto-regressive language model. These include e.g. sentence external information retrieved for a knowledge base or representations of multi-modal encoders.

Overbond, the leading API-based credit trading automation and execution service in the global capital markets, has secured funding from Fitch Ventures, the equity investment arm of Fitch Group, which is a global leader in financial information services. Overbond will use the capital to grow its sales and marketing division with plans to open an office in London, U.K., and double its headcount over the coming year. In addition, through new cloud-based data redistribution channels, Overbond will grow its global presence, integrate new data sources to expand its AI models' coverage and provide enhanced AI trade automation solutions for clients. With the addition of neutrino8 wireless access to the vendor-agnostic .connect Joseph Hospital, along with Perimeter Medical Imaging AI, Inc.("Perimeter" or the "Company") – a medical technology company driven to transform cancer surgery with ultra-high-resolution, real-time, advanced imaging tools to address high unmet medical needs – today jointly announced the first commercial placement of the Perimeter S-Series OCT system in the state of California at Pavilion Surgery Center in Orange, CA.

When the available hardware cannot meet the memory and compute requirements to efficiently train high performing machine learning models, a compromise in either the training quality or the model complexity is needed. In Federated Learning (FL), nodes are orders of magnitude more constrained than traditional servergrade hardware and are often battery powered, severely limiting the sophistication of models that can be trained under this paradigm. While most research has focused on designing better aggregation strategies to improve convergence rates and in alleviating the communication costs of FL, fewer efforts have been devoted to accelerating on-device training. Such stage, which repeats hundreds of times (i.e. In this work, we present the first study on the unique aspects that arise when introducing sparsity at training time in FL workloads. We then propose ZeroFL, a framework that relies on highly sparse operations to accelerate on-device training. Models trained with ZeroFL and 95% sparsity achieve up to 2.3% higher accuracy compared to competitive baselines obtained from adapting a state-of-the-art sparse training framework to the FL setting. Despite it being a relatively new subfield of machine learning (ML), Federated Learning (FL) (McMahan et al., 2017; Reddi et al., 2021; Horvath et al., 2021) has become an indispensable tool to enable privacy-preserving collaboratively learning, as well as to deliver personalised models tailored to the end-user's local data and context (Arivazhagan et al., 2019; Hilmkil et al., 2021; Cheng et al., 2021). Unlike standard centralised training, which normally takes place on the Cloud and makes use of powerful hardware (Hazelwood et al., 2018), FL is envisioned to run on commodity devices such as smartphones or IoT devices often running of batteries, which are orders of magnitude more restricted in terms of compute, memory and power consumption (Qiu et al., 2021). This triplet of factors drastically limits the complexity of the ML models that can be trained on-device in a federated manner, ceiling their usefulness for the aforementioned applications as a result. Other optimization techniques such as quantization and sparsity have been used in the context of FL but mostly as a way to reduce communication costs (Liu et al., 2021; Amiri et al., 2020; Shahid et al., 2021) but not to accelerate on-device training.

Recently has seen the growth in the use of mobile health (mHealth) information services, which have rich guides on improving physical activity. These rich guides evolved from the consideration of various personal behavioural factors, which often deviate from the user's health conditions. The behavioural factors include changing fitness preferences, adherence issues, and uncertainty about future fitness outcomes, which may all lead to a decline in the quality of the mHealth information services. Many of these mHealth information services provide limited fitness guidance owing to the dynamics of the user's health conditions. This paper seeks an adaptive method using deep reinforcement learning to make personalised physical activity recommendations, which is learnt from retrospective physical activity data and can simulate realistic behaviour trajectories. We construct a real-time interaction model for the mHealth information service system based on scientific knowledge about physical activity to evaluate its exercise performance. The physical activity performance evaluation model is used to find the optimal exercise intensity considering the fitness and fatigue effects to avoid the lack of exercise or overload. The short-term activity plans are made using deep reinforcement learning and personal health conditions that change over time. Using this method, we can dynamically update the physical activity recommendation policy in accordance with the real implementation behaviour. Our DRL-based recommender policy was validated by comparison to other benchmark policies. Experimental results show that this adaptive learning algorithm can improve recommendation performance over 4.13 percent.

With the capacity of continual learning, humans can continuously acquire knowledge throughout their lifespan. However, computational systems are not, in general, capable of learning tasks sequentially. This long-standing challenge for deep neural networks (DNNs) is called catastrophic forgetting. Multiple solutions have been proposed to overcome this limitation. This paper makes an in-depth evaluation of the memory replay methods, exploring the efficiency, performance, and scalability of various sampling strategies when selecting replay data. All experiments are conducted on multiple datasets under various domains. Finally, a practical solution for selecting replay methods for various data distributions is provided.