In software engineering, software durability means the solution ability of serviceability of software and to meet user's needs for a relatively long time. Software durability is important for user's satisfaction. For a software security to be durable, it must allow an organization to adjust the software to business needs that are constantly evolving, often in impulsive ways. Durability of software depends on four characteristics mainly; i.e. software trustworthiness, Human Trust for Serviceability, software dependability and software usability.
Fabric computing
Fabric computing or unified computing involves constructing a computing fabric consisting of interconnected nodes that look like a weave or a fabric when seen collectively from a distance. Usually the phrase refers to a consolidated high-performance computing system consisting of loosely coupled storage, networking and parallel processing functions linked by high bandwidth interconnects (such as 10 Gigabit Ethernet and InfiniBand) but the term has also been used to describe platforms such as the Azure Services Platform and grid computing in general (where the common theme is interconnected nodes that appear as a single logical unit). The fundamental components of fabrics are "nodes" (processor(s), memory, and/or peripherals) and "links" (functional connections between nodes). While the term "fabric" has also been used in association with storage area networks and with switched fabric networking, the introduction of compute resources provides a complete "unified" computing system. Other terms used to describe such fabrics include "unified fabric", "data center fabric" and "unified data center fabric". Ian Foster, director of the Computation Institute at the Argonne National Laboratory and University of Chicago suggested in 2007 that grid computing "fabrics" were "poised to become the underpinning for next-generation enterprise IT architectures and be used by a much greater part of many organizations". == History == While the term has been in use since the mid to late 1990s the growth of cloud computing and Cisco's evangelism of unified data center fabrics followed by unified computing (an evolutionary data center architecture whereby blade servers are integrated or unified with supporting network and storage infrastructure) starting March 2009 has renewed interest in the technology. There have been mixed reactions to Cisco's architecture, particularly from rivals who claim that these proprietary systems will lock out other vendors. Analysts claim that this "ambitious new direction" is "a big risk" as companies such as IBM and HP who have previously partnered with Cisco on data center projects (accounting for $2–3bn of Cisco's annual revenue) are now competing with them. In 2007, Wombat Financial Software launched the "Wombat Data Fabric," the first commercial off-the-shelf software platform providing high performance / low-latency RDMA-based messaging across an Infiniband switch. == Key characteristics == The main advantages of fabrics are that massive concurrent processing combined with a huge, tightly coupled address space makes it possible to solve huge computing problems (such as those presented by delivery of cloud computing services); and that they are both scalable and able to be dynamically reconfigured. Challenges include a non-linearly degrading performance curve, whereby adding resources does not linearly increase performance which is a common problem with parallel computing and maintaining security. == Companies == As of 2015 companies offering unified or fabric computing systems include Avaya, Brocade, Cisco, Dell, Egenera, HPE, IBM, Liquid Computing Corporation, TIBCO, Unisys, and Xsigo Systems.
Inception (deep learning architecture)
Inception is a family of convolutional neural network (CNN) for computer vision, introduced by researchers at Google in 2014 as GoogLeNet (later renamed Inception v1). The series was historically important as an early CNN that separates the stem (data ingest), body (data processing), and head (prediction), an architectural design that persists in all modern CNN. == Version history == === Inception v1 === In 2014, a team at Google developed the GoogLeNet architecture, an instance of which won the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The name came from the LeNet of 1998, since both LeNet and GoogLeNet are CNNs. They also called it "Inception" after a "we need to go deeper" internet meme, a phrase from Inception (2010) the film. Because later, more versions were released, the original Inception architecture was renamed again as "Inception v1". The models and the code were released under Apache 2.0 license on GitHub. The Inception v1 architecture is a deep CNN composed of 22 layers. Most of these layers were "Inception modules". The original paper stated that Inception modules are a "logical culmination" of Network in Network and (Arora et al, 2014). Since Inception v1 is deep, it suffered from the vanishing gradient problem. The team solved it by using two "auxiliary classifiers", which are linear-softmax classifiers inserted at 1/3-deep and 2/3-deep within the network, and the loss function is a weighted sum of all three: L = 0.3 L a u x , 1 + 0.3 L a u x , 2 + L r e a l {\displaystyle L=0.3L_{aux,1}+0.3L_{aux,2}+L_{real}} These were removed after training was complete. This was later solved by the ResNet architecture. The architecture consists of three parts stacked on top of one another: The stem (data ingestion): The first few convolutional layers perform data preprocessing to downscale images to a smaller size. The body (data processing): The next many Inception modules perform the bulk of data processing. The head (prediction): The final fully-connected layer and softmax produces a probability distribution for image classification. This structure is used in most modern CNN architectures. === Inception v2 === Inception v2 was released in 2015, in a paper that is more famous for proposing batch normalization. It had 13.6 million parameters. It improves on Inception v1 by adding batch normalization, and removing dropout and local response normalization which they found became unnecessary when batch normalization is used. === Inception v3 === Inception v3 was released in 2016. It improves on Inception v2 by using factorized convolutions. As an example, a single 5×5 convolution can be factored into 3×3 stacked on top of another 3×3. Both has a receptive field of size 5×5. The 5×5 convolution kernel has 25 parameters, compared to just 18 in the factorized version. Thus, the 5×5 convolution is strictly more powerful than the factorized version. However, this power is not necessarily needed. Empirically, the research team found that factorized convolutions help. It also uses a form of dimension-reduction by concatenating the output from a convolutional layer and a pooling layer. As an example, a tensor of size 35 × 35 × 320 {\displaystyle 35\times 35\times 320} can be downscaled by a convolution with stride 2 to 17 × 17 × 320 {\displaystyle 17\times 17\times 320} , and by maxpooling with pool size 2 × 2 {\displaystyle 2\times 2} to 17 × 17 × 320 {\displaystyle 17\times 17\times 320} . These are then concatenated to 17 × 17 × 640 {\displaystyle 17\times 17\times 640} . Other than this, it also removed the lowest auxiliary classifier during training. They found that the auxiliary head worked as a form of regularization. They also proposed label-smoothing regularization in classification. For an image with label c {\displaystyle c} , instead of making the model to predict the probability distribution δ c = ( 0 , 0 , … , 0 , 1 ⏟ c -th entry , 0 , … , 0 ) {\displaystyle \delta _{c}=(0,0,\dots ,0,\underbrace {1} _{c{\text{-th entry}}},0,\dots ,0)} , they made the model predict the smoothed distribution ( 1 − ϵ ) δ c + ϵ / K {\displaystyle (1-\epsilon )\delta _{c}+\epsilon /K} where K {\displaystyle K} is the total number of classes. === Inception v4 === In 2017, the team released Inception v4, Inception ResNet v1, and Inception ResNet v2. Inception v4 is an incremental update with even more factorized convolutions, and other complications that were empirically found to improve benchmarks. Inception ResNet v1 and v2 are both modifications of Inception v4, where residual connections are added to each Inception module, inspired by the ResNet architecture. === Xception === Xception ("Extreme Inception") was published in 2017. It is a linear stack of depthwise separable convolution layers with residual connections. The design was proposed on the hypothesis that in a CNN, the cross-channels correlations and spatial correlations in the feature maps can be entirely decoupled. Training each network took 3 days on 60 K80 GPUs, or approximately 0.5 petaFLOP-days.
Vicuna LLM
Vicuna LLM is an omnibus large language model used in AI research. Its methodology is to enable the public at large to contrast and compare the accuracy of LLMs "in the wild" (an example of citizen science) and to vote on their output; a question-and-answer chat format is used. At the beginning of each round two LLM chatbots from a diverse pool of nine are presented randomly and anonymously, their identities only being revealed upon voting on their answers. The user has the option of either replaying ("regenerating") a round, or beginning an entirely fresh one with new LLMs. (The user also has the option of choosing which LLMs to do battle.) Based on Llama 2, it is an open source project, and it itself has become the subject of academic research in the burgeoning field. A non-commercial, public demo of the Vicuna-13b model is available to access using LMSYS.
Language model benchmark
A language model benchmark is a standardized test designed to evaluate the performance of language models on various natural language processing tasks. These tests are intended for comparing different models' capabilities in areas such as language understanding, generation, and reasoning. Benchmarks generally consist of a dataset and corresponding evaluation metrics. The dataset provides text samples and annotations, while the metrics measure a model's performance on tasks like answering questions, text classification, and machine translation. These benchmarks are developed and maintained by academic institutions, research organizations, and industry players to track progress in the field. In addition to accuracy, the metrics can include throughput, energy efficiency, bias, trust, and sustainability. == Overview == === Types === Benchmarks may be described by the following adjectives, not mutually exclusive: Classical: These tasks are studied in natural language processing, even before the advent of deep learning. Examples include the Penn Treebank for testing syntactic and semantic parsing, as well as bilingual translation benchmarked by BLEU scores. Question answering: These tasks have a text question and a text answer, often multiple-choice. They can be open-book or closed-book. Open-book QA resembles reading comprehension questions, with relevant passages included as annotation in the question, in which the answer appears. Closed-book QA includes no relevant passages. Closed-book QA is also called open-domain question-answering. Before the era of large language models, open-book QA was more common, and understood as testing information retrieval methods. Closed-book QA became common since GPT-2 as a method to measure knowledge stored within model parameters. Omnibus: An omnibus benchmark combines many benchmarks, often previously published. It is intended as an all-in-one benchmarking solution. Reasoning: These tasks are usually in the question-answering format, but are intended to be more difficult than standard question answering. Multimodal: These tasks require processing not only text, but also other modalities, such as images and sound. Examples include OCR and transcription. Agency: These tasks are for a language-model–based software agent that operates a computer for a user, such as editing images, browsing the web, etc. Adversarial: A benchmark is "adversarial" if the items in the benchmark are picked specifically so that certain models do badly on them. Adversarial benchmarks are often constructed after state of the art (SOTA) models have saturated (achieved 100% performance) a benchmark, to renew the benchmark. A benchmark is "adversarial" only at a certain moment in time, since what is adversarial may cease to be adversarial as newer SOTA models appear. Public/Private: A benchmark might be partly or entirely private, meaning that some or all of the questions are not publicly available. The idea is that if a question is publicly available, then it might be used for training, which would be "training on the test set" and invalidate the result of the benchmark. Usually, only the guardians of the benchmark have access to the private subsets, and to score a model on such a benchmark, one must send the model weights, or provide API access, to the guardians. The boundary between a benchmark and a dataset is not sharp. Generally, a dataset contains three "splits": training, test, and validation. Both the test and validation splits are essentially benchmarks. In general, a benchmark is distinguished from a test/validation dataset in that a benchmark is typically intended to be used to measure the performance of many different models that are not trained specifically for doing well on the benchmark, while a test/validation set is intended to be used to measure the performance of models trained specifically on the corresponding training set. In other words, a benchmark may be thought of as a test/validation set without a corresponding training set. Conversely, certain benchmarks may be used as a training set, such as the English Gigaword or the One Billion Word Benchmark, which in modern language is just the negative log-likelihood loss on a pretraining set with 1 billion words. Indeed, the distinction between benchmark and dataset in language models became sharper after the rise of the pretraining paradigm, whereby a model is first trained on massive, unlabeled datasets to learn general language patterns, syntax, and knowledge (pretraining), and the base model is then adapted to specific, downstream tasks using smaller, labeled datasets (fine-tuning). === Lifecycle === Generally, the life cycle of a benchmark consists of the following steps: Inception: A benchmark is published. It can be simply given as a demonstration of the power of a new model (implicitly) that others then picked up as a benchmark, or as a benchmark that others are encouraged to use (explicitly). Growth: More papers and models use the benchmark, and the performance on the benchmark grows. Maturity, degeneration or deprecation: A benchmark may be saturated, after which researchers move on to other benchmarks. Progress on the benchmark may also be neglected as the field moves to focus on other benchmarks. Renewal: A saturated benchmark can be upgraded to make it no longer saturated, allowing further progress. === Construction === Like datasets, benchmarks are typically constructed by several methods, individually or in combination: Web scraping: Ready-made question-answer pairs may be scraped online, such as from websites that teach mathematics and programming. Conversion: Items may be constructed programmatically from scraped web content, such as by blanking out named entities from sentences, and asking the model to fill in the blank. This was used for making the CNN/Daily Mail Reading Comprehension Task. Crowd sourcing: Items may be constructed by paying people to write them, such as on Amazon Mechanical Turk. This was used for making the MCTest. === Evaluation === Generally, benchmarks are fully automated. This limits the questions that can be asked. For example, with mathematical questions, "proving a claim" would be difficult to automatically check, while "calculate an answer with a unique integer answer" would be automatically checkable. With programming tasks, the answer can generally be checked by running unit tests, with an upper limit on runtime. The benchmark scores are of the following kinds: For multiple choice or cloze questions, common scores are accuracy (frequency of correct answer), precision, recall, F1 score, etc. pass@n: The model is given n {\displaystyle n} attempts to solve each problem. If any attempt is correct, the model earns a point. The pass@n score is the model's average score over all problems. k@n: The model makes n {\displaystyle n} attempts to solve each problem, but only k {\displaystyle k} attempts out of them are selected for submission. If any submission is correct, the model earns a point. The k@n score is the model's average score over all problems. cons@n: The model is given n {\displaystyle n} attempts to solve each problem. If the most common answer is correct, the model earns a point. The cons@n score is the model's average score over all problems. Here "cons" stands for "consensus" or "majority voting". The pass@n score can be estimated more accurately by making N > n {\displaystyle N>n} attempts, and use the unbiased estimator 1 − ( N − c n ) ( N n ) {\displaystyle 1-{\frac {\binom {N-c}{n}}{\binom {N}{n}}}} , where c {\displaystyle c} is the number of correct attempts. For less well-formed tasks, where the output can be any sentence, there are the following commonly used scores including BLEU ROUGE, METEOR, NIST, word error rate, LEPOR, CIDEr, and SPICE. === Issues === error: Some benchmark answers may be wrong. ambiguity: Some benchmark questions may be ambiguously worded. subjective: Some benchmark questions may not have an objective answer at all. This problem generally prevents creative writing benchmarks. Similarly, this prevents benchmarking writing proofs in natural language, though benchmarking proofs in a formal language is possible. open-ended: Some benchmark questions may not have a single answer of a fixed size. This problem generally prevents programming benchmarks from using more natural tasks such as "write a program for X", and instead uses tasks such as "write a function that implements specification X". inter-annotator agreement: Some benchmark questions may be not fully objective, such that even people would not agree with 100% on what the answer should be. This is common in natural language processing tasks, such as syntactic annotation. shortcut: Some benchmark questions may be easily solved by an "unintended" shortcut. For example, in the SNLI benchmark, having a negative word like "not" in the second sentence is a strong signal for the "Contradiction" category, regardless of what the se
CPU modes
CPU modes (also called processor modes, CPU states, CPU privilege levels and other names) are operating modes for the central processing unit of most computer architectures that place restrictions on the type and scope of operations that can be performed by instructions being executed by the CPU. For example, this design allows an operating system to run with more privileges than application software by running the operating systems and applications in different modes. Ideally, only highly trusted kernel code is allowed to execute in the unrestricted mode; everything else (including non-supervisory portions of the operating system) runs in a restricted mode and must use a system call (via interrupt) to request the kernel perform on its behalf any operation that could damage or compromise the system, making it impossible for untrusted programs to alter or damage other programs (or the computing system itself). Device drivers are designed to be part of the kernel due to the need for frequent I/O access. Multiple modes can be implemented, e.g. allowing a hypervisor to run multiple operating system supervisors beneath it, which is the basic design of many virtual machine systems available today. == Mode types == The unrestricted mode is often called kernel mode, but many other designations exist (master mode, supervisor mode, privileged mode, etc.). Restricted modes are usually referred to as user modes, but are also known by many other names (slave mode, problem state, etc.). Hypervisor Hypervisor mode is used to support virtualization, allowing the simultaneous operation of multiple operating systems. Kernel and user In kernel mode, the CPU may perform any operation allowed by its architecture; any instruction may be executed, any I/O operation initiated, any area of memory accessed, and so on. In the other CPU modes, certain restrictions on CPU operations are enforced by the hardware. Typically, certain instructions are not permitted (especially those—including I/O operations—that could alter the global state of the machine), some memory areas cannot be accessed, etc. User-mode capabilities of the CPU are typically a subset of those available in kernel mode, but in some cases, such as hardware emulation of non-native architectures, they may be significantly different from those available in standard kernel mode. Some CPU architectures support more modes than those, often with a hierarchy of privileges. These architectures are often said to have ring-based security, wherein the hierarchy of privileges resembles a set of concentric rings, with the kernel mode in the center. Multics hardware was the first significant implementation of ring security, but many other hardware platforms have been designed along similar lines, including the Intel 80286 protected mode, and the IA-64 as well, though it is referred to by a different name in these cases. Mode protection may extend to resources beyond the CPU hardware itself. Hardware registers track the current operating mode of the CPU, but additional virtual-memory registers, page-table entries, and other data may track mode identifiers for other resources. For example, a CPU may be operating in Ring 0 as indicated by a status word in the CPU itself, but every access to memory may additionally be validated against a separate ring number for the virtual-memory segment targeted by the access, and/or against a ring number for the physical page (if any) being targeted. This has been demonstrated with the PSP handheld system. Hardware that meets the Popek and Goldberg virtualization requirements makes writing software to efficiently support a virtual machine much simpler. Such a system can run software that "believes" it is running in supervisor mode, but is actually running in user mode. == Architectures == Several computer systems introduced in the 1960s, such as the IBM System/360, DEC PDP-6/PDP-10, the GE-600/Honeywell 6000 series, and the Burroughs B5000 series and B6500 series, support two CPU modes; a mode that grants full privileges to code running in that mode, and a mode that prevents direct access to input/output devices and some other hardware facilities to code running in that mode. The first mode is referred to by names such as supervisor state (System/360), executive mode (PDP-6/PDP-10), master mode (GE-600 series), control mode (B5000 series), and control state (B6500 series). The second mode is referred to by names such as problem state (System/360), user mode (PDP-6/PDP-10), slave mode (GE-600 series), and normal state (B6500 series); there are multiple non-control modes in the B5000 series. === RISC-V === RISC-V has three main CPU modes: User Mode (U), Supervisor Mode (S), and Machine Mode (M). Virtualization is supported via an orthogonal CSR setting instead of a fourth mode.
Application performance engineering
Application performance engineering is a method to develop and test application performance in various settings, including mobile computing, the cloud, and conventional information technology (IT). == Methodology == According to the American National Institute of Standards and Technology, nearly four out of every five dollars spent on the total cost of ownership of an application is directly attributable to finding and fixing issues post-deployment. A full one-third of this cost could be avoided with better software testing. Application performance engineering attempts to test software before it is published. While practices vary among organizations, the method attempts to emulate the real-world conditions that software in development will confront, including network deployment and access by mobile devices. Techniques include network virtualization.