Statistical measurements as a way to study mode partition in injection lasers

We have demonstrated a statistical measurement technique to characterize the mode partition in injection lasers. In the steady state small nonlasing side modes have distributions similar to that of a narrow-band Gaussian noise. There is a certain probability that they carry a significant portion of the total output. For example, if a side mode of a TJS (AlGa)As laser contains on the average 1.6 percent of the total output power, there is a probability of 10^-6, that it carries 30 percent of the total output. Distributions of main modes of a multimode laser can be matched by beta distributions with properly selected parameters. In addition to dc conditions, this technique has also been used to study the transient behavior of injection lasers. The dominant mode has been shown to start from a noise distribution.     

Efficacy of microwave coagulation therapy (MCT) in patients with liver tumors

We evaluated the efficacy of microwave coagulation therapy (MCT) in 84 patients with hepatocellular carcinomas (HCC) and 40 with metastatic liver tumors (MLT). The response rates calculated with diagnostic imaging were 92% in HCC and 80% in MLT. The regional recurrence rates were relatively higher in patients with MLT (33%) than in HCC (14%). The average surgical margin in operative MCT group was 11 mm. The cumulative survival rates at three and five years were 63% and 38% in HCC and 43% and 33% in MLT, respectively. The complications were similarly encountered in HCC and MLT (12% versus 13%). When these observations are taken together, MCT is a radical and safe locoregional therapy which can keep an adequate surgical margin and assure long survival.     

Automatically Controlled 2-Vessel Pressure-Cooker Test-Equipment

A highly accelerated humidity test for plastic encapsulated IC reliability short-term evaluation has been studied. 2-Vessel pressure cooker test (PCT) equipment capable of controlling relative humidity and temperature independently, and of keeping specimens free from water droplet condensation, has been designed. This equipment consists of a test chamber and a vapor chamber. Humidity conditions are set by controlling the temperature difference between the test chamber and the vapor chamber. Humidity levels are controlled with the direct pressure adjustment. The whole test chamber is heated in an oven, thereby obtaining temperature stability and uniformity. Using this equipment, humidity tests were carried out on plastic molded ICs. As a result, good test reproducibility and excellent correlation in test results between PCT and conventional humidity test such as 85°C/85%RH, were obtained. As a consequence, it was found possible, in a short time, to evaluate plastic molded IC reliability quantitatively in humidity ambients using this 2-vessel PCT equipment. This equipment can be used for quality assurance testing of plastic-encapsulated ICs in a short time.     

Baseband characteristics of long-wavelength l.e.d. systems

As multimode fibres improve, the transfer characteristics of l.e.d.-based systems become dominated by chromatic effects. We derive a simple formula describing the baseband transfer function and verify it by comparative laser and l.e.d. dispersion measurements.     

New electro-optical effect: optical activity of electric-field-induced twisted-nematic liquid crystal

A new electro-optical effect in nematic or nematic-cholesteric mixed liquid crystals is described which allows a 90° polarisation plane rotation of linearly polarised incident light induced by the electric field.     

Subspace information criterion for model selection

The problem of model selection is considerably important for acquiring higher levels of generalization capability in supervised learning. In this article, we propose a new criterion for model selection, the subspace information criterion (SIC), which is a generalization of Mallows's C(L). It is assumed that the learning target function belongs to a specified functional Hilbert space and the generalization error is defined as the Hilbert space squared norm of the difference between the learning result function and target function. SIC gives an unbiased estimate of the generalization error so defined. SIC assumes the availability of an unbiased estimate of the target function and the noise covariance matrix, which are generally unknown. A practical calculation method of SIC for least-mean-squares learning is provided under the assumption that the dimension of the Hilbert space is less than the number of training examples. Finally, computer simulations in two examples show that SIC works well even when the number of training examples is small.