Regulation of the PI-3 kinase (PI3K)/Akt signalling pathway is essential for maintaining the integrity of fundamental cellular processes, cell growth, survival, death and metabolism, and dysregulation of this pathway is implicated in the development and progression of cancers. Receptor tyrosine kinases (RTKs) are major upstream regulators of PI3K/Akt signalling. The phosphatase and tensin homologue (PTEN), a well characterised tumour suppressor, is a prime antagonist of PI3K and therefore a negative regulator of this pathway. Loss or inactivation of PTEN, which occurs in many tumour types, leads to overactivation of RTK/PI3K/Akt signalling driving tumourigenesis. Cellular PTEN levels are tightly regulated by a number of transcriptional, post-transcriptional and post-translational regulatory mechanisms. Of particular interest, transcription of the PTEN pseudogene, PTENP1, produces sense and antisense transcripts that exhibit post-transcriptional and transcriptional modulation of PTEN expression respectively. These additional levels of regulatory complexity governing PTEN expression add to the overall intricacies of the regulation of RTK/PI-3 K/Akt signalling. This review will discuss the regulation of oncogenic PI3K signalling by PTEN (the regulator) with a focus on the modulatory effects of the sense and antisense transcripts of PTENP1 on PTEN expression, and will further explore the potential for new therapeutic opportunities in cancer treatment.

We introduce two tactics, namely the strategicallytimed attack and the enchanting attack, to attack reinforcement learning agents trained by deep reinforcement learning algorithms using adversarial examples. In the strategically-timed attack, the adversary aims at minimizing the agent's reward by only attacking the agent at a small subset of time steps in an episode. Limiting the attack activity to this subset helps prevent detection of the attack by the agent. We propose a novel method to determine when an adversarial example should be crafted and applied. In the enchanting attack, the adversary aims at luring the agent to a designated target state. This is achieved by combining a generative model and a planning algorithm: while the generative model predicts the future states, the planning algorithm generates a preferred sequence of actions for luring the agent. A sequence of adversarial examples is then crafted to lure the agent to take the preferred sequence of actions. We apply the proposed tactics to the agents trained by the state-of-the-art deep reinforcement learning algorithm including DQN and A3C. In 5 Atari games, our strategically-timed attack reduces as much reward as the uniform attack (i.e., attacking at every time step) does by attacking the agent 4 times less often. Our enchanting attack lures the agent toward designated target states with a more than 70% success rate. Example videos are available at http: //yclin.me/adversarial_attack_RL/.

No abstract

Watching a 360 • sports video requires a viewer to continuously select a viewing angle, either through a sequence of mouse clicks or head movements. To relieve the viewer from this "360 piloting" task, we propose "deep 360 pilot" -a deep learning-based agent for piloting through 360 • sports videos automatically. At each frame, the agent observes a panoramic image and has the knowledge of previously selected viewing angles. The task of the agent is to shift the current viewing angle (i.e. action) to the next preferred one (i.e., goal). We propose to directly learn an online policy of the agent from data. Specifically, we leverage a state-of-the-art object detector to propose a few candidate objects of interest (yellow boxes in Fig. 1). Then, a recurrent neural network is used to select the main object (green dash boxes in Fig. 1). Given the main object and previously selected viewing angles, our method regresses a shift in viewing angle to move to the next one. We use the policy gradient technique to jointly train our pipeline, by minimizing: (1) a regression loss measuring the distance between the selected and ground truth viewing angles, (2) a smoothness loss encouraging smooth transition in viewing angle, and (3) maximizing an expected reward of focusing on a foreground object. To evaluate our method, we built a new 360-Sports video dataset consisting of five sports domains. We trained domain-specific agents and achieved the best performance on viewing angle selection accuracy and users' preference compared to [53] and other baselines.

The severe acute respiratory syndrome (SARS) epidemic of 2003 was responsible for 774 deaths and caused significant economic damage worldwide. Since July 2003, a number of SARS cases have occurred in China, raising the possibility of future epidemics. We describe here a rapid, sensitive, and highly efficient assay for the detection of SARS coronavirus (SARS-CoV) in cultured material and a small number (n ‫؍‬ 7) of clinical samples. Using rolling circle amplification (RCA), we were able to achieve sensitive detection levels of SARS-CoV RNA in both solid and liquid phases. The main advantage of RCA is that it can be performed under isothermal conditions with minimal reagents and avoids the generation of false-positive results, a problem that is frequently encountered in PCR-based assays. Furthermore, the RCA technology provides a faster, more sensitive, and economical option to currently available PCR-based methods.Severe acute respiratory syndrome (SARS) is an emerging disease caused by the novel SARS coronavirus (SARS-CoV) (2,4,5,14). By the end of the SARS epidemic in July 2003, a total of 8,096 SARS cases had been reported from 30 countries, with 774 deaths. Whether future outbreaks of SARS will occur is unknown at present. However, given the recent SARS cases in southern China arising from an unknown source and a number of laboratory-related infections (12), it is important to be prepared for such a possibility. In the absence of a SARSCoV vaccine or antiviral drugs, the use of strict infection control policies and early diagnosis with rapid, sensitive, and highly specific laboratory methods are essential for the early management of SARS-CoV infection.Apart from epidemiological linkages, the clinical and radiographic features of the disease are not SARS specific, identifying a need for specific laboratory tests that can confirm SARS-CoV infection early in the course of the illness. Detection of SARS-CoV-specific antibodies is a sensitive and specific but is not possible at clinical presentation (6,14).Detection of SARS-CoV by reverse transcription-PCR (RT-PCR) in clinical specimens allows diagnosis in the early stage of the disease. However, in contrast to many other acute respiratory infections, only low levels of SARS-CoV are thought to be present during the early symptomatic phase of infection. On the basis of the results of first-generation RT-PCR assays, SARS-CoV RNA can be detected with a sensitivity of only ca. 30 to 50% in a single respiratory specimen. A higher sensitivity can be achieved if serial samples are collected, particularly during the second week of illness when maximal virus shedding occurs (13,14). The type of clinical sample (e.g., nasopharyngeal aspirate, throat swabs, stool samples, urine, etc.) also affects the sensitivity of .Recently, the utility of circularizable oligonucleotides, or "padlock probes," has been demonstrated for the detection of target nucleic acid sequences; this approach shows greater sensitivity than conventional PCR (3,8,16). Upon hybridization to a target DNA or...