The Frank-Wolfe (FW) optimization algorithm, due to its projection free property, has gained popularity in recent years with typical application within the field of machine learning. In the stochastic setting, it is still relatively understudied in comparison to the more expensive projected method of Stochastic Gradient Descent (SGD). We propose a novel Stochastic Frank-Wolfe (SFW) algorithm, inspired by SGDo where random sampling is considered without replacement. In analogy to this abbreviation, we call the proposed algorithm SFWo. We consider a convex setting of gradient Lipschitz (smooth) functions over compact domains. Depending on experiment design (LASSO, Matrix Sensing, Matrix Completion), the SFWo algorithm, exhibits a faster or matching empirical convergence, and a tendency of bounded suboptimality by the precursor SFW. Benchmarks on both synthetic and real world data display that SFWo improves on the number of stochastic gradient evaluations needed to achieve the same guarantee as SFW. / Intresset för Frank-Wolfes (FW) optimeringsalgoritm har tack vare dess projektionsfria egenskap ökat de senaste åren med typisk tillämpning inom området maskininlärning. I sitt stokastiska uförande är den fortfarande relativt understuderad i jämförelse med den dyrare projicerande metoden Stochastic Gradient Descent (SGD). Vi föreslår en ny Stochastic Frank-Wolfe(SFW) algoritm, inspirerad av SGDo där slumpmässigt urval görs utan återläggning. I analogi med denna förkortning så kallar vi den föreslagna algoritmen SFWo. Vi betraktar en konvex miljö och gradient Lipschitz kontinuerliga (släta) funktioner över kompakta definitionsmängder. Beroende på experimentdesign (LASSO, Matrix Sensing, Matrix Completion) så visar den föreslagna algoritmen SFWo på en snabbare eller matchande empirisk konvergens och tenderar vara begränsad i suboptimalitet av föregångaren SFW. Prestandajämförelser på både syntetisk och verklig data visar att SFWo förbättrarantalet stokastiska gradientevalueringar som behövs för att uppnå samma garanti som för SFW.

Smoking combustible cigarettes is the major cause of disease and death among adults living in the United States (U.S.). In fact, smoking combustible cigarettes causes nearly half a million premature deaths among U.S. adults every year.1–4 It is estimated that over 14% (equating to 34 million persons) of U.S. adults smoke cigarettes currently.2,5 This is a substantial decrease from 1965 when it was estimated that more than 42% of U.S. adults smoked cigarettes.1 This is partly attributed to the well-established evidence that smoking cigarettes causes harm to almost every human organ system5 and is associated with an elevated risk of developing cancer6, cardiovascular disease 7, pulmonary disease and respiratory illnesses.8–10 Despite the well-established health consequences of smoking cigarettes, millions of people are still smoking, which alone suggests that nicotine (the primary constituent of cigarettes) is highly addictive.3 Thankfully, smoking cessation by means of pharmacological treatments has been shown to help smokers overcome nicotine addiction. A review of the research on the efficacy of varenicline (Chantix) and Nicotine Replacement Therapy (NRT), two of the most commonly used smoking cessation treatments, reveals that both treatments increase long-term smoking abstinence rates with odds ratios of 3.85 for varenicline and 1.74 for NRT when compared to placebo.11–14 Even more, both drugs appear to be generally well-tolerated, with no known life-threatening side effects when compared to placebo. Research shows that the most common side effects for varenicline are nausea, insomnia, gastrointestinal effects, headache and abnormal dreams.11,15 The most common side effects for NRT appear to be skin irritation, insomnia, headache, nausea/vomiting and cough.11,15 Ultimately, both varenicline and NRT appear to be strong options for achieving smoking abstinence both with respect to overall efficacy and tolerability.

Knee osteoarthritis (OA) is a progressive disease that ultimately requires patients to receive a total knee arthroplasty (TKA) to replace the damaged structures within the knee with an artificial joint. Surgeons have many options when selecting an appropriate implant. Patients want a TKA that feels 'normal' and allows them to perform most activities without pain, stiffness, and other residual symptoms. However, 20% of patients remain unsatisfied with their surgery, regardless. This thesis aimed to examine the effect of implant selection during TKA on knee biomechanical function during various ADLs. Several gaps were identified within the review of literature: 1) patient-reported outcome measures cannot differentiate between medial ball and socket (MBS) and posterior stabilized (PS) implants, 2) most biomechanical studies were performed only in postoperative patients, and 3) studies that compared MBS and PS implants were primarily focused on level walking conditions, and overlooked tasks that placed more demand on the knee joint. Twenty-eight individuals with severe knee OA were randomized to receive either an MBS (n=14) or PS implant. They completed a biomechanical assessment within one month and one year after TKA and were compared to 14 controls of similar age, sex, and body mass index. They performed a variety of tasks which explored three main areas: 1) examine the alterations in gait variability among individuals with OA following a TKA procedure using either a PS or MBS implant; 2) enhance the understanding of the post-operative effects of TKA with either MBS and PS implants on knee biomechanics and muscle activities during level walking, as well as more demanding tasks such as descending a ramp or staircase; 3) simulate the dynamic knee joint loads in post-operative TKA patients with either PS or MBS implants during closed-chain, bilateral tasks such as sit-to-stand. Initially, a series of studies were performed to develop a new test called waveform-level variance inequality test (eqvartest), which had not been previously utilized in the literature. This test was used to identify discrepancies in gait variability pre and post-TKA in the gait cycle. Following TKA, patients showed decreased variability in knee moment and power at single-limb support. Neither the MBS nor PS implant provided the same level of variability as the control group, demonstrating reduced knee joint stability. The MBS group had a gait pattern closer to the control group during level walking, whereas the PS group walked with a stiffer knee. However, during more demanding ADLs, the differences were less apparent. During ramp descent, knee joint stability issues became prominent as MBS and PS groups adopted a 'cautious gait pattern,' widening their base of support and stiffening their knee to reduce loading. During stair descent, the MBS implant provided increased stability as it required less muscle activity than the PS, requiring greater hamstring muscle activation. During sit-to-stand, MBS and PS groups favoured their non-operated knee as they had reduced total vertical, medial, and lateral KCF on their operated knee compared to their non-operated side. This may be due to compensatory strategies developed through the progression of knee OA and may increase the risk of developing knee OA on the non-operated limb. The outcomes of this thesis can assist clinicians in selecting the most appropriate implant for their patients and guide them in designing rehabilitation programs that can enhance patient function following TKA.

The increasing speed gap between microprocessors and off-chip DRAM makes last-level caches (LLCs) a critical component for computer performance. Multi core processors aggravate the problem since multiple processor cores compete for the LLC. As a result, LLCs typically consume a significant amount of the die area and effective utilization of LLCs is mandatory for both performance and power efficiency. We present a novel replacement policy for last-level caches (LLCs). The fundamental observation is to view LLCs as a shared resource among multiple address streams with each stream being generated by a static memory access instruction. The management of LLCs in both single-core and multi-core processors can then be modeled as a competition among multiple instructions. In our proposed scheme, we prioritize those instructions based on the number of LLC accesses and reuses and only allow cache lines having high instruction priorities to replace those of low priorities. The hardware support for our proposed replacement policy is light-weighted. Our experimental results based on a set of SPEC 2006 benchmarks show that it achieves significant performance improvement upon the least-recently used (LRU) replacement policy for benchmarks with high numbers of LLC misses. To handle LRU-friendly workloads, the set sampling technique is adopted to retain the benefits from the LRU replacement policy.

A replacement female's value is primarily determined by her reproductive potential and the expected value of calves produced. To improve sales revenues, sellers benefit from understanding the buyers' valuation of physical characteristics related to reproductive potential and calf values. The goal of this research is to identify the impact of physical characteristics on the valuation of individual replacement females through a hedonic pricing model. Results suggest all facets of pregnancy (i.e. pregnancy status, months pregnant, expected due-date, and cow-calf pairs) are crucial to the valuation. Particularly, pregnant replacement females are discounted relative to non-pregnant, ascending in value as months pregnant increases and reaching a premium over non-pregnant status at approximately five months. It is suspected that newly pregnant replacements are discounted due to higher abortion risks. Finally, the largest premiums were observed for cow-calf pairs, where risk of abortion is zero and the replacement female has proven her reproductive potential.

No description available.

No description available.

No description available.

No description available.

No description available.